Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors
Rust
docs: refresh competitor research and roadmap
AkitaOnRails3052eb475e96
diff --git a/docs/comparison.md b/docs/comparison.mdindex 3211b3d9..7634f2aa 100644--- a/docs/comparison.md+++ b/docs/comparison.md@@ -95,46 +95,47 @@ setup.## Maturity and maintenanceThis is a crowded, fast-moving field, and nearly every tool compared here is-actively maintained. As of 2026-10-13, the tools re-checked that day had-commits in October, with two exceptions: Letta's original server is retired-(its work continues in Letta Code), and memvid has had no commits since-2026-07-14. Coding-agent memory now sits at the top of the star chart:+actively maintained. As of 2026-10-18, the tools re-checked that day had+recent October work except for the quiet projects named below. Letta's+original server is retired (its work continues in Letta Code), but its GitHub+repository is an active redirect with the old source on an `archive` branch.+memvid has had no commits since 2026-07-14. Coding-agent memory sits at the top of the star chart:claude-mem leads, MemPalace is third, and most of the app-memory leaders(Mem0, Hindsight, OpenViking, Supermemory, Cognee) ship coding-agent plugins.| Project | Stars (~) | Latest release | Maintenance ||---|---|---|---|-| claude-mem | 100.3k | v13.37.0 (2026-10-13) | active |-| Mem0 | 67.1k | core v2.2.1 (2026-09-25) | active |-| MemPalace | 59.5k | v3.10.0 | active |+| claude-mem | 101.6k | v14.0.1 (2026-10-17) | active |+| Mem0 | 67.4k | core v2.2.1 (2026-09-25) | active |+| MemPalace | 59.6k | v3.11.0 (2026-10-14) | active || Hindsight | 49.1k | v1.0.0 (2026-10-13) | active |-| OpenViking | 39.8k | v0.5.0 (2026-10-09) | active |-| Cognee | 32.2k | v1.6.3 (2026-10-07) | active |-| Zep/Graphiti | 31.7k | v0.30.2 (2026-09-08) | active |-| Supermemory | 31.2k | server-v0.0.9 (2026-10-12) | active |-| agentmemory | 29.4k | v0.9.30 (2026-10-06) | active |-| TencentDB Agent Memory | 28.6k | v2.0.2 (2026-10-09) | active |-| Letta (V1 server) | 25.1k | 0.16.8, frozen | retired (archived 2026-08-16) |-| Letta Code | 3.6k | v0.34.13 (2026-10-13) | active |-| Memori | 17.1k | not re-read | not re-read |-| memvid | 16.6k | not re-read | no commits since 2026-07-14 |-| memU | 14.5k | PyPI `memu-cli` 0.11.0b3 (2026-08-21) | active |-| EverOS | 13.4k | v1.4.1 (2026-09-24) | active |-| MemOS | 11.8k | v2.0.34 | not re-read |-| Honcho | 7.2k | tag v3.2.0 | active (as of 2026-09-18) |-| Engram | 6.8k | v2.0.0 (2026-09-18) | active (as of 2026-09-22) |-| basic-memory | 4.0k | 2026-08-25 | active (as of 2026-09-18) |-| mcp-memory-service | 2.0k | 2026-09-14 | active (as of 2026-09-18) |-| LangMem | 1.7k | PyPI-only | active (as of 2026-09-18) |-| Caura | 0.5k | backend-v3.17.2 (2026-09-19) | active (as of 2026-09-22) |-| LiquidLM | closed-source | `@liquidlm/cli` 0.1.5 (2026-09-17) | active (young, solo; as of 2026-09-22) |+| OpenViking | 40.3k | v0.5.0 (2026-10-09) | active |+| Cognee | 32.5k | v1.6.5 (2026-10-16) | active |+| Zep/Graphiti | 32.0k | v0.31.1 (2026-10-15) | active |+| Supermemory | 31.3k | server-v0.0.9 (2026-10-12) | active |+| agentmemory | 29.7k | v0.9.30 (2026-10-06) | active |+| TencentDB Agent Memory | 29.6k | v2.0.3-beta.1 (2026-10-13) | active |+| Letta (V1 server) | 25.2k | 0.16.8, frozen | retired; source on `archive` branch |+| Letta Code | 3.7k | v0.34.16 (2026-10-16) | active |+| Memori | 17.2k | v3.3.6 (2026-05-28) | quiet since 2026-10-03 |+| memvid | 16.6k | v2.0.140 (2026-05-27) | no commits since 2026-07-14 |+| memU | 14.6k | PyPI `memu-cli` 0.10.0 stable | active |+| EverOS | 13.5k | v1.4.2 (2026-10-14) | active |+| MemOS | 11.9k | v2.0.34 (2026-09-23) | active; no change after cutoff |+| Honcho | 7.7k | v3.3.0 (2026-10-14) | active |+| Engram | 7.4k | v3.4.0 (2026-10-12) | active |+| basic-memory | 4.3k | v0.23.2; v0.24 unreleased | active |+| mcp-memory-service | 2.0k | v12.1.1 (2026-10-16) | active |+| LangMem | 1.7k | PyPI 0.0.30 | quiet |+| Caura | 0.5k | backend-v3.34.5 (2026-10-18) | active |+| LiquidLM | closed-source | `@liquidlm/cli` 0.1.6 stable | active |Adjacent tools, not memory systems: codebase-memory-mcp (~46.5k, a codegraph), qmd (~30.4k, markdown search), context-mode (~27.9k, survivescompaction with a hook-fed per-project SQLite FTS5 store) and beads (~27.8k,an issue tracker with `bd remember`).-Full figures, sources, and per-tool caveats: [`research-2026-landscape.md`](research-2026-landscape.md#popularity-and-maintenance-signal-as-of-2026-10-13).+Full figures, sources, and per-tool caveats: [`research-2026-landscape.md`](research-2026-landscape.md#popularity-and-maintenance-signal-as-of-2026-10-18).## How the field validates the approach@@ -228,8 +229,11 @@ ships something ai-memory does not, it is named.of SQLite + Chroma, the same memory under 20+ harnesses, claim-oncehandoffs, cross-project messaging, and per-project multi-user auth withouta paid cloud tier. You give up its to-do tools, the bridge that answers- Claude Code's and Codex's native memory-folder reads from its store, and- the past observations it shows when the agent opens a file.+ Claude Code's and Codex's native memory-folder reads from its store, the+ past observations it shows when the agent opens a file, and v14's polished+ local “My Work” dashboard. TeamBrain adds optional paid shared-work and+ collision views; ai-memory shares project memory without that cloud layer+ but has no equivalent human work queue yet.- **From MemPalace:** you keep hook capture for Claude Code, Codex and Cursorand a shared server. You gain git-versioned markdown pages as the sourceof truth, claim-once handoffs, typed cross-project messaging, and the auth@@ -359,7 +363,10 @@ ships something ai-memory does not, it is named.embeddings). You give up (for now) memU's turnkey skill-distillationpackaging, its `memu-agent detect` for unknown agents, and its hosted cloud,and EverOS's ANN vector index, optional Milvus backend and published- end-to-end benchmark.+ end-to-end benchmark. EverOS v1.4.2 also reserves query-embedding capacity+ from background writes; ai-memory's reranker fails open under saturation,+ but embedding backfill and foreground query embedding do not yet have an+ equivalent measured capacity split.- **From Caura (governed multi-agent fleet memory):** a different buyer. Cauraoptimizes org-scale fleet governance (agent/team/org visibility scopes, fourtrust tiers, an audit log, PII flagging) on Postgres+pgvector with LLM@@ -369,7 +376,10 @@ ships something ai-memory does not, it is named.cross-fleet trust governance across hundreds of agents, use Caura; if you wantfile-first, zero-LLM, single-binary project memory, use ai-memory. (Caura'strust-tier governance is a real capability we do not match; see- [`competitive-parity.md`](competitive-parity.md).)+ [`competitive-parity.md`](competitive-parity.md).) Caura's beta+ collaboration extension also adds live peer discovery, presence, leases,+ correlated replies, and runtime wake helpers. That is fleet orchestration,+ not durable project memory, and external effects remain at-least-once.- **From TencentDB Agent Memory (proxy fan-in server):** you trade a Nodethree-service proxy that intercepts each harness's API base URL (distillingChat Memory / Skill / Wiki / CodeGraph assets) for a file-first binary that@@ -382,17 +392,24 @@ ships something ai-memory does not, it is named.## Where we're behind, or different by choice-- **Distribution through plugin directories.** Cognee's Claude Code plugin is- listed in Anthropic's official plugin directory, and Mem0, Supermemory and- OpenViking ship native plugins per agent. ai-memory- installs with `install.sh` and `ai-memory setup` (3.0) or `install-hooks`,- not from a plugin marketplace.-- **An installer that detects and wires every agent.** `ai-memory setup`- (3.0) wires MCP, hooks and Agent Skills for each detected agent on Linux and- macOS; 3.1 brought the same two-command install and a per-user service step to- Windows (`install.ps1` + a logon entry that starts `ai-memory service-host`).- Hindsight's coding-agent package covers 16+ agents, and memU's `memu-agent- detect` finds unknown agents' transcripts.+- **No human workstream operations view.** Managed workstreams, run leases,+ handoffs, and portable event history exist, but `/web` does not turn them+ into a scannable queue of active, blocked, review-ready, and completed work.+ claude-mem v14's local “My Work” view is better for a human supervising+ several requests.+- **No structured page filters on search.** Scope and `as_of` are typed, but+ `memory_query` cannot narrow by path prefix, kind, tier, tag, or author+ before ranking. Mem0 has directory/category scopes; basic-memory's+ unreleased v0.24 adds a broader structured filter set.+- **Foreground embedding has no measured reservation from backfill.** The LLM+ reranker is bounded and fails open, but query embedding and administrative+ embedding backfill share the configured embedder without a dedicated+ foreground budget. EverOS v1.4.2 shows the latency value of separating them.+- **Marketplace reach is narrower than harness support.** ai-memory now ships+ its own Claude Code and Codex marketplace plugins and `ai-memory setup`+ detects and wires installed agents across Linux, macOS, and Windows. It is+ not listed in Anthropic's official directory, and long-tail agents still use+ setup/MCP recipes where no native plugin format exists.- **No end-to-end QA number.** We publish retrieval hit@5; Mem0, Zep,Hindsight and EverOS publish end-to-end QA accuracy with a reader and ajudge model. Our opt-in `answer=true` synthesis (2.4) has only an early,diff --git a/docs/competitive-parity.md b/docs/competitive-parity.mdindex a878c591..f6b15993 100644--- a/docs/competitive-parity.md+++ b/docs/competitive-parity.md@@ -1,6 +1,6 @@# Competitive parity & migration-worthiness-A self-critical internal audit, September 2026, refreshed 2026-10-13 with+A self-critical internal audit, September 2026, refreshed 2026-10-18 withthe research and gap decisions below. Companion to the public[`comparison.md`](comparison.md) (the fair pitch) and[`research-2026-landscape.md`](research-2026-landscape.md) (what each competitor@@ -55,17 +55,17 @@ None of the competitors below has more than one or two of these; none has all.## Migration verdicts, by competitor-Refreshed 2026-10-13. Coding-agent memory is now the most-starred part of the+Refreshed 2026-10-18. Coding-agent memory is now the most-starred part of thefield, and several app-memory products (Mem0, Supermemory, Hindsight,OpenViking, Cognee) ship coding-agent plugins, so they appear here ascoding-agent competitors as well as in their original camps.| From | Migrate to ai-memory? | You gain | You give up ||---|---|---|---|-| **claude-mem** (hook capture, LLM observer) | **Yes** for zero-LLM, multi-harness or team use; **not obviously** for a solo Claude Code user happy with it | Zero-LLM capture (it compresses every observation with an LLM), git-markdown truth instead of SQLite FTS5 + Chroma, claim-once handoffs, cross-project messaging, per-project multi-user auth without a paid cloud tier | Its to-do tools, three-step progressive search, the bridge that answers Claude Code's and Codex's native memory-folder reads from its store, and past observations surfaced when the agent reads a file |+| **claude-mem** (hook capture, LLM observer) | **Yes** for zero-LLM, multi-harness or team use; **not obviously** for a solo Claude Code user happy with it | Zero-LLM capture (it compresses every observation with an LLM), git-markdown truth instead of SQLite FTS5 + Chroma, claim-once handoffs, cross-project messaging, per-project multi-user auth without a paid cloud tier | Its to-do tools, three-step progressive search, native-memory bridge, file observations, and v14's polished local “My Work” queue; TeamBrain adds paid shared-work/collision views || **MemPalace** (hook capture, team server) | **Partly**: overlapping goals, different substrate | Git-markdown truth, claim-once handoffs, typed cross-project messaging, the auth ladder | Its temporal graph and team server as shipped. Its 96.6% / 98.4% R@5 is retrieval-only like our hit@5, but on a different harness, so neither number ranks the tools || **mcp-memory-service** (fact rows + hook capture) | **Yes** if you liked hook capture + typed edges + published numbers | Files-as-truth (not opaque fact rows), handoffs, cross-project messaging, `as_of` temporal, a reproducible number, and (new on 2.4) its aging set at parity done zero-LLM/reversible/off-by-default (per-tier decay curves, extractive compression, DBSCAN cold-cluster dedup, access boosts, 0.4–0.75-band contradiction flagging) | Cloud/multi-backend replication (Cloudflare/Milvus), graph viz. (Belief-strength + clustering consolidation are no longer a give-up: ai-memory now ships both, clustering as a zero-LLM default-off pass, belief-strength off and eval-gated) |-| **basic-memory** (file-first markdown) | **Yes** for multi-harness coding continuity; **no** for a personal Obsidian KB with live team co-edit | Ambient capture (no `write_note` ceremony), full lifecycle, handoffs/messaging, git-versioned truth, published numbers | Local zero-cost reranking, real-time collaborative editing, Postgres backend, `build_context` graph-walk, hosted mobile/web UX |+| **basic-memory** (file-first markdown) | **Yes** for multi-harness coding continuity; **no** for a personal Obsidian KB with live team co-edit | Ambient capture (no `write_note` ceremony), full lifecycle, handoffs/messaging, git-versioned truth, published numbers, native OKF v0.2, conditional writes, temporal reads | Local zero-cost reranking, real-time collaborative editing, Postgres backend, `build_context` graph-walk, hosted mobile/web UX, document extraction, and broader structured filters. The latter two are in its unreleased v0.24 changelog, not its stable release || **agentmemory** (hook capture, KV store) | **Yes** for operability + data ownership | Self-contained binary (no `iii-engine` sidecar, no ELv2 runtime), files-as-truth, SQL indexes committed in-txn, zero-LLM summaries and handoffs, typed claim-once handoffs, a multi-user auth ladder (DB tokens, OIDC), native Windows without a manual sidecar download (3.1 `install.ps1` + `setup`), and a 3.0 importer for its vault and `ExportData`. Both now default to auth on for new installs (theirs since 0.9.30, ours in 3.0) | Browser edit/forget (the 3.0 web UI adds page history/diff, a session timeline, a recall trace and a health panel, but stays read-only by design), the iii console's OTEL traces (we ship explain timings + Prometheus `/admin/metrics` in 3.0), a P2P mesh-sync primitive (`ai-memory-wikisync` is our git-based answer), the opt-in local cross-encoder reranker, an npm/`npx` install (ours is a two-command native install on Linux, macOS and Windows in 3.0/3.1, no npm package), ~31 more tools (54, all visible by default; mostly speculative). Benchmarks are not like-for-like: their 95.2% LongMemEval-S is a component measurement with the reranker disabled (their 2026-10-09 baseline adds 84.68% strict recall-all@5, which they say supports no competitive claim), ours is end-to-end || **Claude Code built-in memory** | **Yes** for teams / multi-machine / multi-harness; **not obviously** for a solo dev on one machine in Claude only | Cross-machine, cross-harness, team sharing, tool-lifecycle capture, real search | Zero setup: it's already on with no server to run (our one structural disadvantage for the solo case; 3.0's `install.sh` + `ai-memory setup` cut ours to two commands on Linux and macOS, but there is still a server) || **Hindsight** (pages-over-facts; coding-agent package) | **Yes** for zero-LLM, files-as-truth, or claim-once handoffs; **no** if you want its end-to-end QA numbers, git-history ingestion, or 16+ agent installer | Zero-LLM default (Hindsight needs an LLM, local or by subscription), markdown files as the source of truth (its markdown is a one-way mirror of Postgres), one binary, claim-once handoffs, cross-project messaging, and (new on 2.4) a belief-strength `confidence` + an opt-in "dream" rewrite that borrow their model but never delete a source and keep the zero-LLM default | Its published end-to-end QA accuracy, git-history ingestion, hooks for 16+ coding agents with one shared bank per repo, a reranker on by default, image/file ingestion, user/team/company scopes inside a bank, and belief-strength that *actually moves ranking by default* (ours ships but is off and eval-gated) |@@ -78,9 +78,9 @@ coding-agent competitors as well as in their original camps.| **Honcho** (Plastic Labs) | **No: different problem.** It models the *human* (theory-of-mind peer representations); we remember the *project* | If you (mis)used it for coding continuity: file-first git truth, zero-LLM, single-binary self-host, per-project scoping, no mandatory LLM egress | Its user-psychology reasoning engine (Deriver/Dreamer/Dialectic), published user-recall benchmarks, managed cloud, none of which a coding-memory user needs || **LiquidLM** (solo-built hosted memory API, Supermemory camp) | **Weak: different job.** It's a cloud "second brain" with multimodal RAG; we're automatic, file-first coding memory | Data ownership (markdown+git, no opaque cloud), zero-LLM default, single-binary self-host, automatic hook capture, per-project team sharing, no API/subscription spend | Its multimodal ingestion (video/audio/PDF/Office), polished web app + grounded chat, managed hosting, GitHub-sync connector, and out-of-the-box rerank || **Engram** (self-hosted coding-agent store, issue #810) | **Yes**, same lane | Files-as-truth git-markdown wiki (vs SQLite-as-truth + git-synced compressed chunks), **automatic lifecycle-hook capture** (vs agent-driven `mem_*` MCP saves), typed claim-once handoffs + cross-project messaging, multi-user page sharing, published numbers | Engram's zero-config Go binary is equally lean; you give up its managed Cloud replication add-on. A near-tie on posture; the switch is about capture model + wiki-vs-rows, not moat depth |-| **EverOS** (file-first, LLM-leaning, issue #810) | **Yes** for zero-LLM/single-binary/team; **no** if you want its vector tier, user-track modeling, or end-to-end QA numbers | Zero-LLM default (EverOS's minimal tier still needs an LLM for the core flow), one self-contained binary (vs a Python stack), cross-harness lifecycle capture, typed handoffs/messaging | EverOS's ANN vector index and optional Milvus backend, its published end-to-end benchmarks (LoCoMo 94.42, LongMemEval-S 94.00, with exact models recorded per result file), its first-class user `episodes/profile` track, and its packaged offline "memory evolution" UX |+| **EverOS** (file-first, LLM-leaning, issue #810) | **Yes** for zero-LLM/single-binary/team; **no** if you want its vector tier, user-track modeling, or end-to-end QA numbers | Zero-LLM default (EverOS's minimal tier still needs an LLM for the core flow), one self-contained binary (vs a Python stack), cross-harness lifecycle capture, typed handoffs/messaging | EverOS's ANN vector index and optional Milvus backend, its published end-to-end benchmarks (LoCoMo 94.42, LongMemEval-S 94.00), its first-class user `episodes/profile` track, packaged offline “memory evolution,” and v1.4.2's explicit foreground/background embedding-capacity split || **memU** (LLM-wiki + skill distillation, issue #810) | **Partly**: same file-first instinct, different default | Zero-LLM capture/retrieval by default (memU's distill flow is LLM-required), one binary (vs Python + sidecars patching host instruction files), MCP + hooks, git-versioned truth | memU's turnkey **skill-distillation-into-Markdown** packaging, `memu-agent detect` for unknown agents, and its hosted cloud (memu.so). Our nearest equivalent is the opt-in experience/auto-improve pass, not a one-command skill distiller |-| **Caura** (governed fleet memory, issue #810) | **No: different buyer** (org-scale multi-agent governance) | If you (mis)used it for one project: file-first git truth, zero-LLM, single-binary self-host, per-project scoping, no LLM egress | Its **fleet-governance layer** (org/team/agent visibility scopes and **four cross-fleet trust tiers**), which ai-memory does **not** match (see the gaps below); plus its managed platform and self-reported LoCoMo/LongMemEval numbers |+| **Caura** (governed fleet memory, issue #810) | **No: different buyer** (org-scale multi-agent governance and live peer orchestration) | If you (mis)used it for one project: file-first git truth, zero-LLM, single-binary self-host, per-project scoping, no LLM egress | Its **fleet-governance layer** (org/team/agent visibility scopes and four trust tiers) plus the beta bus's peer discovery, presence, leases, correlated replies, and runtime wake helpers; ai-memory intentionally does not run a live agent fleet || **TencentDB Agent Memory** (proxy fan-in server, issue #810) | **Mostly no: different capture model** | Hook/MCP capture into a file-first wiki you own (vs a base-URL proxy intercepting traffic), zero-LLM default (its v2 hub needs an LLM), one binary vs a Node three-service stack, typed handoffs | Its zero-code proxy integration, the v2 hub's per-asset access control, and its **CodeGraph** codebase index (adjacent code-intelligence we don't do). Note: needs no Tencent DB (SQLite default); the name misleads || **Mem0/Zep app SDKs; Supermemory/OpenViking hosted products; LiquidLM, Honcho, Caura** | different buyer (app/user personalization, hosted context DB / RAG-as-a-service, user-modeling, org-scale fleet governance), **not migration targets** for those products; their coding plugins are covered above | — | — |@@ -88,8 +88,8 @@ Net: for a **self-hosted, multi-harness, team, zero-LLM, data-ownership**coding-memory use case, ai-memory does the basics and adds things none of themfully have, so migration is justified. The gap has narrowed: hook capture andgit-backed markdown are now common, and several tools ship things we do not-(plugin-directory distribution, git-history ingestion, recall on every-prompt, end-to-end QA numbers). The premise does *not* hold for app/user+(official-directory reach, git-history ingestion, recall on every prompt,+human work queues, end-to-end QA numbers). The premise does *not* hold for app/userpersonalization (Mem0/Supermemory/LiquidLM), user-modeling /theory-of-mind (Honcho), enterprise graph queries (Zep Cloud), agent-buildingruntimes (Letta), or a solo dev happy with Claude's built-in or with a@@ -101,7 +101,9 @@ rather than overclaim.The main finding: ai-memory did **not** cargo-cult this field. Of the ideas ittook, most are at parity-or-better; the exceptions are **scoped-but-unfinishedsubstrate**, each deliberately gated behind "prove it with an eval first." The-remaining risk is that borrowed ideas sit **dormant** until the eval harness (R2) that would justify activating them gets built.+remaining risk is that opt-in ideas stay **dormant** because the retrieval+eval has not shown enough benefit to justify activating them by default; the+next evidence gap is reproducible end-to-end QA.| Borrowed idea | From | Verdict | Detail ||---|---|---|---|@@ -127,10 +129,11 @@ remaining risk is that borrowed ideas sit **dormant** until the eval harness (R2| LLM merge/rewrite "dream" pass | Hindsight (Dreamer) / Supermemory ("dreaming") | **SHIPPED, opt-in/off (parity of mechanism, unproven)** | 2.4 (#816) rewrites cold clusters into one page via the gated `apply_batch` path, `dry_run` first, JSON-schema output; **never deletes a source** (superseded with a merge-note stub, `restore-page` recovers), off by default, provider-less stores keep the zero-LLM A3 path. Off until an R2 quality delta is shown || bi-temporal-lite | Zep/Graphiti | **BEHIND by design** | Ingestion-time only; world-time validity deferred (because dating facts needs an LLM and breaks the zero-LLM path) |-## Honest gaps (consolidated) and recommended adjustments+## Prior gaps and current status-These are documented recommendations only. **Nothing here is scheduled or to be-implemented without a separate decision.** Ordered by leverage.+This ledger records what earlier audits asked for, what shipped, and what is+still open. Open items remain recommendations until a separate decision+schedules them.1. **R2 reproducible eval harness, built (#771/#772) and run.** It was themeta-blocker: belief-strength ranking, typed-edge rank weighting, and a@@ -219,20 +222,19 @@ implemented without a separate decision.** Ordered by leverage.belief-strength confidence, and the LLM dream pass), audited row-by-row inthe borrowed-ideas table above; the design of record is[`design-memory-aging.md`](design-memory-aging.md).- MCP tool behavior hints now ship across the 23-tool surface (title plus+ MCP tool behavior hints now ship across the 24-tool surface (title plusread-only/destructive/idempotent/open-world), with the runtime auth andscope checks remaining authoritative because annotations are untrustedmetadata. **Still unshipped:** typed-edge **rank** weighting and a`contradicts` authority cap (A5 flags, it doesn't yet rank);- a read-only graph visualization in `/web`; a first-class TS SDK story for- ecosystem parity.+ a read-only graph visualization in `/web`.Deliberately out of scope: Honcho's theory-of-mind user-modeling engine andLiquidLM's opaque-cloud / LLM-mandatory / multimodal-second-brain substrate(all off-mission for file-first, zero-LLM coding memory; the global-scopepreferences page already covers the useful user-prefs slice).7. **Fix doc staleness the audit surfaced** (accuracy, not features):- `research-basic-memory.md` understates basic-memory's shipped cross-encoder- reranking + Teams tier. The agentmemory items were corrected on 2026-10-12:+ `research-basic-memory.md` now distinguishes stable v0.23.2 from the+ broader unreleased v0.24 changelog. The agentmemory items were corrected on 2026-10-12:`comparison.md`, `benchmarks/README.md` and `evals/README.md` no longerattribute the `0.967` (the grep baseline on their 15-query`coding-agent-life-v1` corpus) to LongMemEval-S, where they publish 95.2%@@ -253,7 +255,10 @@ implemented without a separate decision.** Ordered by leverage.its banks carry user/team/company scopes, and its LLM can be local;EverOS publishes benchmarks and an MCP server and was listed twice (asEverMemOS); memU's releases ship on PyPI; and the "smaller, less- VC-amplified niche" positioning no longer holds.+ VC-amplified niche" positioning no longer holds. The 2026-10-18 pass then+ corrected Letta's retirement wording, current releases, ai-memory's own+ marketplace/install coverage, and the recommendations that had already+ shipped.8. **Fleet-scale trust governance: a real gap vs Caura (issue #810).** Ourmulti-user story is per-project (`pages shared, batons owned`, invariant #16)with a root→DB-user→OIDC auth ladder and an audit log, and it is deliberately@@ -275,6 +280,10 @@ invariants), **ALREADY** (we have it and should say so).| Capability seen in the field | Verdict | Why ||---|---|---|+| Human workstream operations view (claude-mem v14 “My Work”) | ADOPT | The managed-workstream ledger already has runs, leases, events, and handoff evidence, but humans cannot scan active/attention/review-ready work in `/web`. Build a read-only derived view; add only a minimal review marker, not a second task system |+| Foreground query capacity isolated from embedding backfill (EverOS v1.4.2) | MEASURE-FIRST, then ADOPT if contention is real | The reranker already has a shared fail-open gate. Query embedding and administrative backfill share the embedder without a foreground reservation. Add a contention benchmark; preserve lexical/entity/graph results under saturation |+| Structured pre-rank filters (Mem0; basic-memory v0.24 unreleased) | ADOPT | Bounded path-prefix, kind, tier, tag, author, and valid-time filters improve monorepo/tool ergonomics without changing ranking. All paths must reuse `ScopeResolver` and fail closed on incomplete scope |+| Per-operation LLM provider/model profiles (Graphiti v0.31; Hindsight `main`) | MEASURE-FIRST | Potentially lowers cost and latency for consolidation, rerank, and auto-improvement, but expands config and auth surface. Prototype against the evals before adding public settings || Retrieval recall through capture (several tools rank whole raw sessions) | SHIPPED (3.0) | The 2026-10-12 decomposition showed our zero-LLM session page indexed only 80-character turn titles; the observation stream recovered it (FTS hit@5 0.662 → 0.951, local 0.800 → 0.945, all 470). Chasing another tool's headline number is declined: those numbers measure different things || Local zero-LLM cross-encoder reranker | MEASURE-FIRST | Fits the zero-LLM invariant (a local model, not an LLM). Bar before building: hit@5 at least +0.03 on both FTS and local paths over the shipped ranking, p95 at most +100 ms on the reference host and measured on a low-end ARM host; default-on only if the keyless path gains. Measured headroom is small after the stream (component-mode FTS 0.972 vs pipeline 0.951) || Editable browser viewer (edit / forget / bulk forget) | DECLINE | The web UI is read-only by design (`frontend-api.md` §10); wiki mutations go only through `Wiki::write_page`/`apply_batch`; agents already edit and forget via MCP (`memory_write_page` with `expected_page_id`, `memory_delete_page`) and the CLI |@@ -302,26 +311,28 @@ invariants), **ALREADY** (we have it and should say so).### Step up, borrow, ignore-From the 2026-10-13 research. Each line names the tools that ship the+From the 2026-10-18 research. Each line names the tools that ship thecapability. Items marked *verify first* depend on our own current behavior,which should be checked before any work is scoped.**Step up** (gaps a coding-agent user notices):-1. **Plugin-directory listing and one-command marketplace installs** for- Claude Code and Codex (Cognee is listed in Anthropic's plugin directory;- Mem0, Supermemory and OpenViking ship native plugins).-2. **An installer that detects and wires every installed agent** (Hindsight- covers 16+ agents; memU's `memu-agent detect` finds unknown ones).- *Verify first*: how much of this `ai-memory setup` already covers.-3. **One reproducible end-to-end QA number** beside hit@5, with reader,+1. **A human workstream operations view** over existing managed runs, with+ attention/review states and bounded evidence (claude-mem v14). This should+ expose work already tracked, not create a kanban subsystem.+2. **Structured filters before ranking** for path prefix, kind, tier, tag,+ author, and valid time (Mem0; basic-memory v0.24, currently unreleased).+3. **Foreground embedding isolation backed by a contention benchmark**+ (EverOS v1.4.2). Reserve query capacity only if the benchmark shows+ administrative backfill can harm interactive latency.+4. **One reproducible end-to-end QA number** beside hit@5, with reader,judge, context tokens and latency stated, and more benchmarks thanLongMemEval (Mem0, Zep, Hindsight and EverOS publish QA numbers; Hindsight'sAMB harness is open, so an adapter is worth considering). Most numberspublished in the field are QA numbers.-4. **Git-history ingestion on first run**, zero-LLM and sanitized+5. **Git-history ingestion on first run**, zero-LLM and sanitized(Hindsight), so a new repository starts with its commit history as context.-5. **Answer harness-native memory reads from our store** (claude-mem's bridge+6. **Answer harness-native memory reads from our store** (claude-mem's bridgefor Claude Code and Codex memory folders). Careful: this touches trustand routing, and the gap decisions above decline a bidirectional bridge.@@ -344,6 +355,9 @@ undelivered event for a later drain.- An opt-in local cross-encoder reranker (Hindsight, basic-memory,agentmemory), gated on the recall-eval bar in the table above.- Per-step model routing (Graphiti's `LLMRuntime`): cheaper consolidation.+- A dry-run-first unified audit/repair command (MemPalace v3.11): consolidate+ the checks already spread across `doctor`, lint, reindex, health, and backup+ only if one operator workflow is demonstrably simpler.- A read-only MCP mode per connection (OpenViking gateway, hosted Zep):safer shared servers.- Opt-in recall on every prompt (Supermemory, OpenViking gateway).@@ -362,9 +376,10 @@ undelivered event for a later drain.- Past observations for a file when the agent reads it (claude-mem).**Ignore** (off-mission or conflicts with an invariant): user profiles,-data connectors, graph/ontology queries, kanban boards, proxy capture,+data connectors in core, graph/ontology queries, generic kanban boards, proxy capture,mounted filesystems, harness-owned self-editing memory, multimodal memory,-hosted clouds, ANN indexes, temporal graphs, and to-do tools. Per-asset ACLs+hosted clouds, ANN indexes, temporal graphs, live fleet presence/wake buses,+and to-do tools. Per-asset ACLs(TencentDB v2, OpenViking) are also out of scope but recorded as a real gapnext to Caura's trust tiers.diff --git a/docs/research-2026-landscape.md b/docs/research-2026-landscape.mdindex 6f3bd624..294154c6 100644--- a/docs/research-2026-landscape.md+++ b/docs/research-2026-landscape.md@@ -1,4 +1,4 @@-# Agent-Memory Landscape, September 2026 - Research Report+# Agent-Memory Landscape, October 2026 - Research Report> Follow-up to the May 2026 research pass (`research-agentmemory.md`,> `research-basic-memory.md`, `research-cognee.md`,@@ -9,13 +9,11 @@> change. Sources at the end; every load-bearing claim was checked against> a live page in September 2026, not remembered.>-> **Refreshed 2026-10-13.** The popularity table, the camp table and the-> entries for Mem0, OpenViking, Zep/Graphiti, Cognee, Supermemory,-> agentmemory, TencentDB Agent Memory, Letta, Hindsight, memU and EverOS were-> re-read against primary sources (repositories, release metadata from the-> GitHub API, vendor research pages) on that date. Claude-mem, MemPalace,-> Memori, MemOS and memvid were added. Corrections are made in place;-> [the 2026-10-13 refresh](#refresh-2026-10-13-what-changed) in §3 summarizes what moved.+> **Refreshed 2026-10-18.** Every tracked competitor was checked against its+> official repository, releases, package metadata, or product documentation.+> The [2026-10-18 delta ledger](#refresh-2026-10-18-one-by-one-delta) records+> both material changes and explicit no-change findings. Older corrections+> remain in place so the report keeps its audit trail.## 1. The headline: our architectural bet got standardized@@ -29,10 +27,10 @@ memory servers advertise themselves as "OKF-backed", and the format ispositioned as the interop layer that makes agent memory portable acrosstools.-Our `wiki/` **is nearly an OKF bundle already**: markdown, YAML-frontmatter, one page per concept, typed via our own frontmatter-conventions. The delta is a conventions mapping, not an architecture-change.+Our `wiki/` is now a native **OKF v0.2 bundle**: markdown, YAML frontmatter,+one page per concept, with generated `index.md` and `log.md` projections and+an `ai-memory okf check` conformance command. The interoperability work is+already shipped.The second validation came from an unexpected direction: **Letta**(formerly MemGPT) published "Is a Filesystem All You Need?" arguing that@@ -76,7 +74,10 @@ gate that installs the packaged artifacts and runs capture and recovery.**basic-memory** is alive, with the same positioning (structured markdown bothhumans and LLMs edit; local files). No architectural surprises. Its-niche overlaps most with OKF, which may absorb its differentiation.+niche overlaps most with OKF, which may absorb its differentiation. Its+unreleased v0.24 changelog now describes native OKF v0.2, optimistic writes,+structured and cross-project filters, document extraction, temporal+qualifiers, a local reranker, locked notes, and cloud team collaboration.**cognee** has moved into content-led growth ("best memoryframework" listicles on its own blog rank for every comparison query)@@ -352,8 +353,9 @@ grounded, cited chat. It is on **every axis we deliberately chose against**: the**vault is an opaque hosted store** (no git, no markdown, no self-host, nothing to`grep`/diff), **LLM is intrinsic and not optional** (Gemini-class extraction/embeddings, no zero-LLM mode), and the memory engine is **closed-source** (the-only open artifact is the MIT npm client `@liquidlm/cli`, v0.1.5 published-2026-09-17; the GitHub `cli` repo is archived and empty). Where it is+only open artifact is the MIT npm client `@liquidlm/cli`, stable v0.1.6, with+v0.2.0-alpha.1 as a prerelease; the GitHub `cli` repo is archived and empty).+Where it isinteresting is **breadth of ingestion and a clean MCP tool surface**: heavymultimodal capture (video/audio/image/PDF/Office, auto-transcription, a`yt-dlp`-fetch-locally pattern), a GitHub repo-sync connector, and eight native@@ -411,8 +413,11 @@ which is our consolidation/auto-improve loop under other names. It validates thebut from the LLM-leaning corner: the minimal tier still **requires an LLM for thecore flow** (embedding/reranking optional, OpenRouter the default; a zero-LLMdemo exists but is not the operating default), it is a Python stack rather than a-single binary. *Update 2026-10-13 (v1.4.1, ~13.4K stars):* native Windows-support, an ANN vector index and an optional Milvus backend. An official+single binary. *Update 2026-10-18 (v1.4.2, ~13.5K stars):* native Windows+support, an ANN vector index and an optional Milvus backend. v1.4.2 also+separates foreground query embedding from background writes, batches+embeddings, self-repairs index damage, and hardens full-scope keys and path+containment. An official`everos-mcp` server (EverMind-AI/EverOS-MCP) was documented on 2026-10-12,so the earlier "MCP support is not stated" no longer holds. Benchmarks arenow published: `benchmarks/README.md` (2026-09-08) reports LoCoMo 94.42 and@@ -587,20 +592,22 @@ with another or with ours, and no ranking is claimed.| Cognee | BEAM 0.79 (20 questions, one conversation), 0.67 at 10M | End-to-end LLM-judged QA | Tuned and scored on the same questions; gpt-5 for ingestion, answering and judging || memvid | +35% LoCoMo (relative) | Relative claim | Unverified |-**claude-mem (`thedotmack/claude-mem`, ~100.3K stars, Apache-2.0, v13.37.0-on 2026-10-13).** Five lifecycle hooks feed SQLite FTS5 plus Chroma, and an+**claude-mem (`thedotmack/claude-mem`, ~101.6K stars, Apache-2.0, v14.0.1+on 2026-10-17).** Five lifecycle hooks feed SQLite FTS5 plus Chroma, and anLLM "observer" compresses every observation, so capture is LLM-driven. Apaid cloud tier (CMEM Pro) exists, and the README endorses a cryptocurrencytoken. Recent work: non-blocking hooks that write to a spool, to-do tools, athree-step progressive search, a bridge that intercepts reads of ClaudeCode's and Codex's native memory folders and answers them from its ownstore, and surfacing past observations for a file when the agent reads it.-No benchmarks are published.+v14 adds the local “My Work” request/workstream dashboard and the optional+paid TeamBrain shared-work layer. No benchmarks are published.-**MemPalace (v3.10.0, ~59.5K stars).** See §2 for the benchmark history.+**MemPalace (v3.11.0, ~59.6K stars).** See §2 for the benchmark history.It now ships hooks for Claude Code, Codex and Cursor, a team server and atemporal graph, and its README publishes retrieval-only numbers (96.6% R@5-raw, 98.4% on 450 held-out questions).+raw, 98.4% on 450 held-out questions). v3.11 adds dry-run-first audit/repair+and optional multimodal EmbeddingGemma 2 retrieval.**Mem0 (~67.1K stars, core v2.2.1 on 2026-09-25).** Native coding-agentplugins for Claude Code, Codex, Cursor, Kimi, Antigravity, OpenCode and pi@@ -621,11 +628,10 @@ April 2026 retrieval is a single-pass add-only extraction plus semantic +BM25 + entity fusion (the OSS package needs the `[nlp]` extra), so the old"vector-only" description is wrong.-**Zep / Graphiti (~31.7K stars, v0.30.2 on 2026-09-08).** v0.30.0 sends+**Zep / Graphiti (~32.0K stars, v0.31.1 on 2026-10-15).** v0.30.0 sendsNeo4j queries to the configured database and reframes the README as atemporal knowledge graph framework (#1801); v0.30.2 uses a driver per-request and isolates concurrent groups (#1699). Unreleased on `main`-(bumped to 0.31.0): #1964 from Zep's internal fork (combined extraction per+request and isolates concurrent groups (#1699). v0.31 released #1964 from Zep's internal fork (combined extraction perepisode in bulk, date rules for facts, stricter contradiction rules), #1778`LLMRuntime` per-prompt model routing, and an MCP server on MCP SDK 2.x(#1921). FalkorDBLite is an embedded option; Kuzu support is deprecated@@ -633,10 +639,11 @@ because upstream is unmaintained. Hosted Zep has a Context MCP server withper-user OIDC sign-in, read-only connections set by an admin, and a writekill switch; its plugin says it is "not positioned for Claude Code, Codex".-**Letta (`letta-ai/letta` ~25.1K stars; `letta-ai/letta-code` ~3.6K).** The-V1 server is retired: commit #3430 (2026-08-16) archived it at 0.16.8, and+**Letta (`letta-ai/letta` ~25.2K stars; `letta-ai/letta-code` ~3.7K).** The+V1 server is retired: commit #3430 (2026-08-16) moved its source to an+`archive` branch at 0.16.8, andthe README points to Letta Code, a coding agent whose default backend is-Letta Cloud (v0.34.13 on 2026-10-13). Since February 2026 ("context+Letta Cloud (v0.34.16 on 2026-10-16). Since February 2026 ("contextrepositories") every agent's memory is a git repository of Markdown fileswith YAML frontmatter and a `MEMORY.md` index; "dreaming" subagents write inparallel through git worktrees, with no vector index by default. Its essay@@ -646,8 +653,8 @@ a Letta key and keeps one memory across all projects by default. Letta Codealso has a `/doctor`-style memory audit that includes the token cost of thesession brief. The ADE references in older notes describe the retired line.-**Memori (~17.1K stars)** sits in the app-memory camp; its paper-(arXiv:2603.19935) reports LoCoMo 87%. **MemOS (~11.8K stars, v2.0.34)**+**Memori (~17.2K stars)** sits in the app-memory camp; its paper+(arXiv:2603.19935) reports LoCoMo 87%. **MemOS (~11.9K stars, v2.0.34)**belongs to the memory-OS camp and reports LoCoMo 88.83 and LongMemEval89.20 on its own harness. **memvid (~16.6K stars)** has had no commits since2026-07-14 and publishes only relative claims (+35% LoCoMo), unverified.@@ -657,41 +664,80 @@ graph), qmd (~30.4K, markdown search), context-mode (~27.9K, survivescompaction with a per-project SQLite FTS5 store fed by hooks) and beads(~27.8K, an issue tracker with `bd remember`).-### Popularity and maintenance signal (as of 2026-10-13)+### Refresh, 2026-10-18: one-by-one delta-Star counts, releases and last commits come from the GitHub API on-2026-10-13; the "Sept" column is the 2026-09-18 figure where this report-tracked it. Rows marked *not re-read* keep their September figures. Of the-rows re-read with a commit date, all but memvid and the retired Letta server-had commits in October 2026; Letta's original server is retired in favour of-Letta Code.+This pass used the 2026-10-13 refresh as its baseline. `Released` means the+capability is in a tagged release; `main` means it is visible in the default+branch but is not yet a stable release. The table records quiet projects+explicitly.++| Project | Change since 2026-10-13 | What it means for ai-memory |+|---|---|---|+| **claude-mem** | **Released v14.0.0/v14.0.1.** “My Work” turns captured requests into a local dashboard with workstream cards, evidence drill-down, live attention states, and an explicit human “Checked” state. TeamBrain adds selected shared work, collision awareness, decisions, and delegation as a paid service. | The local work dashboard is the clearest new product gap. Build a read-only operational view over ai-memory's existing managed-workstream ledger, with derived attention states and a minimal human review marker. Do not copy the paid cloud or create a second task tracker. |+| **Mem0** | No material memory change. The repository added event promotion; the latest core release remains v2.2.1. | No action. Its category and cloud-first tradeoffs are unchanged. |+| **MemPalace** | **Released v3.11.0.** Adds a dry-run-first audit/repair workflow, optional EmbeddingGemma 2 multimodal retrieval, embedder identity checks, large-palace performance work, and security fixes. v3.12 work is on `main`, not released. | ai-memory already records embedding identity, rejects stale vectors, and has `doctor`, lint, reindex, backup, and health tools. A unified audit UX is worth evaluating; native image/audio/video memory remains outside the text-only core. |+| **Hindsight** | No new stable release after v1.0.0. `main` adds decision-model triage before consolidation, decision-routed reflection/ranking, scheduled full mental-model refreshes, operation-specific request extensions, more providers, and richer source exports. | Measure per-operation model profiles and a cheap structured decision tier before adopting either. Any periodic full refresh must produce reviewed proposals, never silently rewrite source truth. |+| **OpenViking** | No new stable release after v0.5.0. `main` adds mixed image/text queries, scoped one-use downloads, Redis-backed idle commit scheduling, multi-replica account stores, privacy commands, and bot steering. | Most changes target a multimodal distributed context platform. The resource drill-down pattern validates ai-memory's progressive disclosure; the scaling and opaque-media stack does not fit the single-binary markdown core. |+| **Cognee** | **Released v1.6.5.** Adds code docstrings as graph documents, Memify APIs, batched trace processing, base64 embeddings, a Calendar connector, and session/dataset binding. | No new core work. Code graphs and connectors remain adjacent; batched ingestion reinforces the existing single-writer batching design. |+| **Zep/Graphiti** | **Released v0.31.0/v0.31.1.** Per-prompt `LLMRuntime` routing, combined extraction, temporal/dedup fixes, Decisions clients, and MCP SDK/group routing moved from unreleased work into a stable release. | This strengthens the case for a measured per-operation provider/model profile. It does not justify a graph database or LLM-mandatory path. |+| **Supermemory** | No new release after server-v0.0.9. `main` aligns the v5 API/docs and hardens MCP authorization, writable defaults, and deletion by ID. | No product gap. The authorization fixes validate ai-memory's fail-closed scope resolver and capability gates. |+| **agentmemory** | No new release after v0.9.30. `main` contains redaction, vector/compression recovery, bounded graph-index work, Redis mutation serialization, and Devin integration preparation. | No action. ai-memory already has Devin support, bounded capture, recovery, and a stricter sanitizer boundary. |+| **TencentDB Agent Memory** | v2.0.3-beta.1 remains the newest release; subsequent commits are fixes. | No material change to the proxy-fan-in comparison. |+| **Letta V1 server** | No product work. The repository is **not GitHub-archived**: its README redirects active development to Letta Code, while the retired V1 source lives on an `archive` branch. | Correct the old “repository archived” wording. The V1 server remains retired, but the repository itself is an active redirect. |+| **Letta Code** | **Released v0.34.14-v0.34.16.** Adds steering, resuming, and stopping remote Claude Code/Codex subagents, vault inheritance, listener heartbeats, and recovery fixes. | This is an agent runtime feature, not a memory-system gap. Managed workstreams remain harness-neutral. |+| **Memori** | No commits since 2026-10-03 and no new release. | No material change. |+| **memvid** | No commits since 2026-07-14 and no new release. | Remains quiet; its relative benchmark claim is still not independently useful. |+| **memU** | No feature release. Four Windows path/idempotency fixes landed. | No material change. |+| **EverOS** | **Released v1.4.2.** Fixes an FD leak, separates query embedding capacity from background writes, batches embeddings, keys indexes by the full scope tuple, self-repairs damaged indexes, and hardens path containment. | The capacity split is worth copying in principle. ai-memory's reranker already fails open behind a shared gate, but query embedding and administrative backfill share the embedder without a foreground reservation. Add a contention benchmark before changing it. |+| **MemOS** | No commits or release after the prior cutoff. Its current headline remains DeepSeek Harness integration, multimodal memory, memory cubes, asynchronous ingestion, and feedback/correction. | No material change. The memory-OS and multimodal tradeoffs remain different from ai-memory's scope. |+| **Honcho** | **Released v3.3.0.** Adds an optional Chroma backend, scoped custom instructions, cursor pagination, prompt-budget handling, and performance fixes; stable object `uid` work followed on `main`. | No core work. Its theory-of-mind buyer remains different. Operation-scoped instructions are a useful reference if per-operation LLM profiles are designed. |+| **Engram** | No new tag after v3.4.0. `main` adds local aliases/routing preview, project provenance, compact search/history clients, capture guards, tombstones, and sync isolation. | No new gap. ai-memory already has stronger repository identity, routing diagnostics, tombstone/supersession semantics, and conflict-aware wiki sync. |+| **basic-memory** | No new stable release after v0.23.2. The unreleased v0.24 changelog describes optimistic checksummed writes, native OKF v0.2 export/check, cross-project and structured filters, document extraction with provenance, temporal qualifiers, a local reranker, locked notes, more harnesses, and cloud team collaboration. | ai-memory already ships native OKF, conditional writes, temporal queries, attribution, and lifecycle capture. Structured filters are a real gap. Document ingestion belongs in an optional importer; local reranking stays measure-first; live co-edit/cloud is a different product. |+| **mcp-memory-service** | **Released v12.1.0/v12.1.1.** Security fixes include stored-XSS handling and graph-write scoping; automatic supersession is now conservative. `main` adds Codex hooks, store-partitioned SSE, tombstone protection, and BM25 fixes. | No new feature work. The release validates conservative supersession, output encoding, and strict write scoping already required by ai-memory's invariants. |+| **LangMem** | No repository or package release since the prior cutoff. | No material change. |+| **Caura** | **Released backend-v3.34.5.** `main` also carries a beta collaboration extension: tenant-scoped peer discovery, requests, correlated replies, presence, leases, progress, and recovery for running Claude/Codex hosts. External effects remain at-least-once. | Treat the bus as adjacent fleet orchestration. ai-memory messaging is durable project-to-project continuity, not a live peer runtime. Do not add presence or wake orchestration to the memory core. |+| **LiquidLM** | Stable CLI is v0.1.6; v0.2.0-alpha.1 is prerelease. Public client work adds scoped tokens, MCP installation, safer upload reconciliation, search/upload commands, and local media fetching; the engine remains closed. | Correct the version. Its managed multimodal second-brain UX remains outside ai-memory's core. |+| **Claude Code native memory** | Releases 2.1.299-2.1.302 improve compaction and runtime stability but add no material project-memory capability. | No change to the comparison: excellent zero-setup single-harness memory, weak portability and team continuity. |++The adjacent codebase-memory, search, context-compaction, and issue-tracking+projects were not promoted into the competitor set. They remain useful+integration references, but they solve a different primary job.++### Popularity and maintenance signal (as of 2026-10-18)++Star counts, releases and last commits come from official GitHub or package+metadata checked on 2026-10-18; the "Sept" column preserves the 2026-09-18+figure where this report tracked one. Memori, memvid, LangMem, and MemOS had+no commit after the prior cutoff. Letta's original server remains retired in+favour of Letta Code, but its repository is a redirect rather than an archived+GitHub repository.| Project | Stars (~) | Sept | Last commit | Latest release | Status ||---|---|---|---|---|---|-| claude-mem | 100.3k | not tracked | 2026-10-13 | v13.37.0 (2026-10-13) | ACTIVE |-| Mem0 | 67.1k | 65.6k | 2026-10-12 | core v2.2.1 (2026-09-25) | ACTIVE |-| MemPalace | 59.5k | not tracked | not re-read (active) | v3.10.0 | ACTIVE |+| claude-mem | 101.6k | not tracked | 2026-10-18 | v14.0.1 (2026-10-17) | ACTIVE |+| Mem0 | 67.4k | 65.6k | 2026-10-17 | core v2.2.1 (2026-09-25); TS v3.3.1 | ACTIVE |+| MemPalace | 59.6k | not tracked | 2026-10-18 | v3.11.0 (2026-10-14) | ACTIVE || Hindsight | 49.1k | 23.9k | 2026-10-13 | v1.0.0 (2026-10-13) | ACTIVE |-| OpenViking | 39.8k | 38.0k | 2026-10-13 | v0.5.0 (2026-10-09) | ACTIVE |-| Cognee | 32.2k | 30.8k | 2026-10-08 | v1.6.3 (2026-10-07) | ACTIVE |-| Zep/Graphiti | 31.7k | 31.0k | 2026-10-09 | v0.30.2 (2026-09-08); `main` at 0.31.0, unreleased | ACTIVE |-| Supermemory | 31.2k | 30.1k | 2026-10-13 | server-v0.0.9 (2026-10-12) | ACTIVE |-| agentmemory | 29.4k | 29.3k | 2026-10-13 | v0.9.30 (2026-10-06) | ACTIVE |-| TencentDB Agent Memory | 28.6k | 27.2k | 2026-10-13 | v2.0.2 (2026-10-09); v2.0.3-beta.1 (2026-10-13) | ACTIVE |-| Letta (V1 server) | 25.1k | 24.8k | 2026-09-10 (docs only) | 0.16.8, frozen | RETIRED (archived 2026-08-16) |-| Letta Code | 3.6k | not tracked | release 2026-10-13 | v0.34.13 (2026-10-13) | ACTIVE |-| Memori | 17.1k | not tracked | not re-read | not re-read | not re-read |-| memvid | 16.6k | not tracked | 2026-07-14 | not re-read | QUIET (no commits since 2026-07-14) |-| memU | 14.5k | 14.4k | 2026-10-13 | PyPI `memu-cli` 0.10.0 (2026-08-05), 0.11.0b3 (2026-08-21) | ACTIVE |-| EverOS | 13.4k | 13.1k | not re-read | v1.4.1 (2026-09-24) | ACTIVE |-| MemOS | 11.8k | not tracked | not re-read | v2.0.34 | not re-read |-| Honcho | 7.2k | 7.2k | 2026-09-18 | tag v3.2.0 (no GH release) | not re-read |-| Engram | 6.8k | 6.8k | 2026-09-22 | v2.0.0 (2026-09-18) | not re-read |-| basic-memory | 4.0k | 4.0k | 2026-09-16 | v0.23.2 (2026-08-25) | not re-read |-| mcp-memory-service | 2.0k | 2.0k | 2026-09-18 | v11.12.0 (2026-09-14) | not re-read |-| LangMem | 1.7k | 1.7k | 2026-09-09 | none (PyPI-versioned) | not re-read |-| Caura | 0.5k | 0.5k | 2026-09-22 | backend-v3.17.2 (2026-09-19) | not re-read |-| LiquidLM | n/a (closed) | n/a | n/a (engine closed) | `@liquidlm/cli` 0.1.5 (2026-09-17) | not re-read |+| OpenViking | 40.3k | 38.0k | 2026-10-18 | v0.5.0 (2026-10-09) | ACTIVE |+| Cognee | 32.5k | 30.8k | 2026-10-16 | v1.6.5 (2026-10-16) | ACTIVE |+| Zep/Graphiti | 32.0k | 31.0k | 2026-10-17 | v0.31.1 (2026-10-15) | ACTIVE |+| Supermemory | 31.3k | 30.1k | 2026-10-18 | server-v0.0.9 (2026-10-12) | ACTIVE |+| agentmemory | 29.7k | 29.3k | 2026-10-18 | v0.9.30 (2026-10-06) | ACTIVE |+| TencentDB Agent Memory | 29.6k | 27.2k | 2026-10-15 | v2.0.3-beta.1 (2026-10-13) | ACTIVE |+| Letta (V1 server) | 25.2k | 24.8k | no V1 code after retirement | 0.16.8, frozen | RETIRED (source on `archive` branch) |+| Letta Code | 3.7k | not tracked | 2026-10-18 | v0.34.16 (2026-10-16) | ACTIVE |+| Memori | 17.2k | not tracked | 2026-10-03 | v3.3.6 (2026-05-28) | QUIET |+| memvid | 16.6k | not tracked | 2026-07-14 | v2.0.140 (2026-05-27) | QUIET |+| memU | 14.6k | 14.4k | 2026-10-17 | PyPI `memu-cli` 0.10.0 (stable) | ACTIVE |+| EverOS | 13.5k | 13.1k | 2026-10-16 | v1.4.2 (2026-10-14) | ACTIVE |+| MemOS | 11.9k | not tracked | no change after cutoff | v2.0.34 (2026-09-23) | ACTIVE |+| Honcho | 7.7k | 7.2k | 2026-10-17 | v3.3.0 (2026-10-14) | ACTIVE |+| Engram | 7.4k | 6.8k | 2026-10-18 | v3.4.0 (2026-10-12) | ACTIVE |+| basic-memory | 4.3k | 4.0k | 2026-10-18 | v0.23.2 (2026-08-25); v0.24 unreleased | ACTIVE |+| mcp-memory-service | 2.0k | 2.0k | 2026-10-18 | v12.1.1 (2026-10-16) | ACTIVE |+| LangMem | 1.7k | 1.7k | 2026-10-02 | PyPI 0.0.30 (2025-10-27) | QUIET |+| Caura | 0.5k | 0.5k | 2026-10-18 | backend-v3.34.5 (2026-10-18) | ACTIVE |+| LiquidLM | n/a (closed) | n/a | n/a (engine closed) | `@liquidlm/cli` 0.1.6; 0.2.0-alpha.1 prerelease | ACTIVE |Adjacent, not memory systems (not ranked above): codebase-memory-mcp ~46.5k,qmd ~30.4k, context-mode ~27.9k, beads ~27.8k.@@ -706,8 +752,7 @@ PyPI rather than GitHub releases (an earlier "releases lag code" reading herewas a misread); memvid has had no commits since 2026-07-14;LangMem and Honcho publish no GitHub *releases* (PyPI/tag-only versioning);and LiquidLM is **closed-source**, so it has no repo signal at all; its only-public, open artifact is the MIT npm client `@liquidlm/cli` (v0.1.5,-2026-09-17).+public, open artifact is the MIT npm client `@liquidlm/cli` (stable v0.1.6).## 4. Research developments worth knowing@@ -724,23 +769,25 @@ public, open artifact is the MIT npm client `@liquidlm/cli` (v0.1.5,the field's evolution as Storage (preserve trajectories) → Reflection(refine them) → Experience (abstract *across* trajectories, proactive).We are solidly through Storage and Reflection (capture, consolidation,- curator, lint, feedback-driven salience). The frontier it names,- cross-trajectory abstraction, is where our auto-improve is only a- beginning.+ curator, lint, feedback-driven salience). The opt-in experience pass now+ covers cross-trajectory abstraction, although its quality gain remains an+ evaluation question rather than a settled default.- **LongMemEval-V2 (arXiv:2605.12493)**: the benchmark generation hasmoved from "recall the fact" to "behave like an experienced colleague"over long horizons, and finds every evaluated system far from it. Two- implications: our product framing matches where evaluation is going,- and the eval-harness recommendation from the May research (#8 in- `research-agentmemory.md`, never implemented) is now more pressing:- the field publishes audited numbers and we publish none.+ implications: our product framing matches where evaluation is going, and+ the in-repo evaluation harness should keep expanding beyond retrieval. We+ now publish reproducible retrieval-through-capture numbers, but still lack+ a comparable end-to-end QA result with reader, judge, context, and latency+ recorded together.- **Bi-temporal validity (Zep paper, arXiv:2501.13956, now mainstreamvia Graphiti)** is the one *specific* mechanism from the graph campworth studying: every fact/edge carries event-time andingestion-time, so "what did we believe on June 1" and "what was trueon June 1" are both answerable, and corrections are supersessions- rather than overwrites. Our pages already supersede; our `links` and- `entities` rows do not carry validity intervals.+ rather than overwrites. ai-memory now ships ingestion-time temporal+ validity and `as_of` reads on SQLite; true world-time extraction remains a+ deliberate LLM-dependent boundary.- Also active, lower relevance for us: MIRIX (multi-agent memoryspecialization), MemOS (scheduling memory like an OS),A-TMA (state-aware memory failure taxonomy), sleep-time compute@@ -756,146 +803,97 @@ public, open artifact is the MIT npm client `@liquidlm/cli` (v0.1.5,> improving, and is ai-memory worth migrating *to*? It is the self-critical> record behind this section's recommendations.-The May research led us to build: versioned supersession, retention+The May research led us to build versioned supersession, retentionformulas, hybrid RRF retrieval, opt-in LLM consolidation, handoffs as aprotocol, single self-contained binary, typed scope isolation. Every one-of those choices looks *better* in September than it did in May. The+of those choices still fits the field in October. Thesubstrate competitors struggled (mempalace's metaphor, agentmemory'sKV), the file-first camp got standardized, and pages-over-facts got apaper. Nothing in this pass argues for re-architecture. The-recommendations are additive and ranked:--**R1: OKF conformance (small, high leverage).** Map our frontmatter-conventions onto OKF v0.1 (`type` field plus its small vocabulary) and-add an OKF export, possibly just documentation plus a thin-`ai-memory export --okf` view of `wiki/`. Being the *server-grade*-implementation of the format Google standardized helps our positioning:-interop with every OKF-aware consumer, at conventions-mapping cost.-Evaluate whether native conformance (wiki pages *are* OKF files) beats-an export step; the closer to native, the stronger the story.--**R2: Reproducible eval harness against LongMemEval-V2 (medium).** The-oldest open recommendation, now with a sharper reason: benchmark claims-are being audited (mempalace) and honest numbers are being published-(mcp-memory-service). We cannot say where we stand, and "experienced-colleague over long horizons" is our pitch. Harness in-repo,-runnable on demand like `writer_throughput`, numbers in docs with the-run command, and never a marketing claim without the harness. Hindsight-(`research-hindsight.md`) sharpens this further: a funded competitor now-ships a *paper* with per-category LongMemEval sub-scores (91.4% then, 94.6%-on its own harness as of 2026-10), so the-bar is no longer "publish a number" but "publish a harness and a metric-someone else can re-run". Fix the split and the metric (theirs is-accuracy, agentmemory/mcp-memory-service quote R@5; pick one and state-it) so our result is comparable, not just present. **Now load-bearing on-2.4:** the two aging features that can move quality, belief-strength folded-into ranking (#815) and the LLM "dream" pass (#816), ship **off by default-and gated on this harness before they may default on**; no R2 delta has-shown a gain yet, so both remain opt-in and unproven (see-[`design-memory-aging.md`](design-memory-aging.md) R2 acceptance).--**R3: Typed relation edges (small-medium).** `causes` / `fixes` /-`contradicts` on our existing `links`/`entities` model, from the-doobidoo playbook. Cheap because the tables exist; valuable because-`contradicts` feeds the lint pass we already run, and `fixes` makes-bug-page chains retrievable as chains.--**R4: Temporal validity on entities/links (medium, evaluate first).**-Bi-temporal-lite: `valid_from` / `superseded_at` on entity-page links so-"what was the database choice in June" is answerable. Pages already-supersede; this extends the same idea one level down. Worth a design-doc before any code; the Graphiti paper is the reference.--**R5: Cross-session abstraction pass (medium-large, the "Experience"-gap).** A periodic consolidation that reads *across* recent sessions-per project and rewrites pattern/preference pages: the cross-trajectory-abstraction both the survey and TriMem's narrative layer point at.-auto-improve is the natural host; today it is per-session-triggered.-Hindsight's **mental-model tier + reflect loop** (`research-hindsight.md`)-is the shipped reference implementation for exactly this (a background-job that rewrites standing-answer pages across accumulated evidence, that-an agent boots from), and it is worth reading before designing ours. Its-consolidation-as-belief-strength (evidence count + quotes, strengthen/-weaken/extend) also points at how R3/R4 could carry *confidence*, not-just supersession order.--**R6: Local embeddings via ONNX (medium; already reserved).**-`models/` has been reserved for exactly this since M9.5 planning.-Competitors ship all-MiniLM locally by default; it removes the provider-dependency from vector search and makes hybrid retrieval a zero-config-default rather than an opt-in. The `ort` crate is the known path.--**R7: Progressive-disclosure retrieval/brief tiers (medium; evaluate-first, design doc before code).** From OpenViking's L0/L1/L2 model, whose-reported 34-91% input-token cut is almost entirely deferring full content.-We already have the pieces: `memory_query` returns ~24-word snippets (an-L0), `memory_read_page` returns the full body (an L2), and V61 abstract-embeddings + the session brief exist. The gap is an explicit **L1-"overview" tier** and, more concretely, a **tiered on-start brief** that-leads with per-page abstracts and expands to full bodies only on demand-instead of loading core-page bodies up front (`render_session_brief`-today spends its whole char budget on L2). This is a token-efficiency-win on the exact hot path (every opted-in session start) with no new-storage: reuse the abstract already embedded at V61. Worth a design doc-that decides where the L1 overview comes from (frontmatter summary,-first-paragraph extraction, or the existing abstract) before any code;-it composes with R5 (the cross-session abstraction pass produces good-L1 summaries) rather than competing with it. **Partly realized on 2.4:**-extractive tier-down (A2, #808) already persists the L0 abstract + L1-summary + L2 keep-token shape for cold episodic pages-([`design-memory-aging.md`](design-memory-aging.md)). The *tiered on-start-brief* landed after 3.0-([`design-tiered-session-brief.md`](design-tiered-session-brief.md)): every-core page gets an extractive L0 excerpt (`abstract:`, else `summary:`, else-the prose it opens with), and the leftover budget buys L1 excerpts and whole-bodies, pinned and `_slots/` pages first, with crowded `_rules/` excerpts-under one heading (Letta-code MemFS's directory deferral, done at render-time). It needs no LLM, no new tool and no migration. Directory-scoped retrieval-(OpenViking's subtree filters, which replaced its TrieHI recursion in-v0.4.23) is the lower-priority half: a path-prefix filter-on `memory_query` within a project, worth noting but not scheduling-until R7's tiering lands.--**R8: Queryable metadata/tag filters + a published eval triple (small--medium; from Supermemory).** Two borrowable ideas from the entrant that leads-on ergonomics. First, **structured metadata/tag filtering on `memory_query`**:-we already key every row by the `(workspace, project, path)` 3-tuple, but-Supermemory's `containerTags` + arbitrary metadata filters let a caller narrow-recall by attributes (kind, tier, author, tag) *before* ranking. A bounded,-optional filter argument on `memory_query`, composing with FTS5/RRF rather than-replacing it, closes a real query-ergonomics gap at low cost; most of the-plumbing (page frontmatter, kinds/tiers, `_slots`) already exists. Second,-**publish an accuracy / latency / context-tokens triple** from the in-repo-`recall_eval.rs`, mirroring Supermemory's open **MemScore / MemoryBench**-philosophy: honest engineering discipline and the right answer to a field now-competing on self-reported single numbers. It composes with R2 (the-reproducible-harness recommendation) rather than competing. A per-project or-per-user **"profile" page** auto-surfaced on start (Supermemory's standing-profile) is a lighter third idea that maps onto our durable pages + the session-brief, worth noting but ranked below the two above. Deliberately *not* borrowed from this-camp: managed cloud connectors and LLM-required extraction in the **core**. If-connectors ever land they belong in the companion importer as opt-in, outside-the trusted boundary, so the file-first / zero-LLM model is preserved (the same-line drawn for multimodal/PDF import: OCR/transcribe to markdown pages, never an-opaque store). On the other two axes: **security, nothing-to borrow** (Supermemory is weaker on every axis we chose: an opaque default store,-cloud-first, LLM-required, `containerTag` isolation vs our 3-tuple + invariant-#16, so this pass is confirmation of our posture, not a source of borrowings;-the only security-relevant note is defensive: any future importer connector must-cross the existing sanitizer boundary), and **performance, only f16/-half-precision embedding storage is worth *evaluating*** (a stored-format change-with a real recall-vs-size tradeoff, gated on `recall_eval.rs`), while their-"sub-300ms p50" is a publish-a-number contract, not a change we need. Ranked, the-only thing worth actually building is the metadata/tag filter; the eval triple is-worth doing because it de-risks measuring anything else; f16 is evaluate-only;-the rest is hold-or-reject. No implementation is scheduled here.+original recommendations below are retained as an implementation audit;+their headings state what has shipped.++**R1: OKF conformance - shipped.** The wiki is a native OKF v0.2 bundle,+with generated index/log projections and `ai-memory okf check`. No export-only+compatibility layer is needed.++**R2: Reproducible eval harness - shipped for retrieval, still open for+end-to-end QA.** The in-repo harness publishes retrieval-through-capture and+component results with run commands and receipts. It also keeps the+belief-strength and LLM dream features off by default because neither has+shown a useful retrieval delta. The remaining work is a reproducible+end-to-end QA path that records the answerer, judge, context budget, latency,+and exact metric.++**R3: Typed relation edges - shipped.** `causes`, `fixes`, and `contradicts`+are stored and explained, related-page reads can traverse them, and+contradiction lint uses them. Edge kind does not yet affect rank by default.++**R4: Temporal validity - shipped at the evidence-backed boundary.**+Ingestion-time validity, supersession, and `as_of` reads cover historical+belief without requiring an LLM to invent world-time dates. Full bi-temporal+fact extraction remains deliberately out of scope for the zero-LLM path.++**R5: Cross-session abstraction pass - shipped, opt-in.** The experience pass+requires evidence from multiple sessions and feeds reviewed changes through+the auto-improvement path. It remains opt-in because the current eval does not+show enough quality benefit to make it a default.++**R6: Local embeddings - shipped.** A checksum-pinned pure-Rust candle path+provides local embeddings without ONNX or a provider key, with identity stored+beside every vector and stale models excluded.++**R7: Progressive-disclosure brief tiers - shipped.** The session brief now+starts with bounded extracts for all standing pages, then spends the remaining+budget on pinned and slot content. It runs without an LLM or migration. The+remaining retrieval gap is a path-prefix filter, now part of P2 below.++**R8: Queryable metadata/tag filters + a published eval triple - partly+open.** The retrieval harness and accuracy/latency/context reporting exist.+Search still cannot filter by path, kind, tier, author, or tag before ranking.+That bounded filter surface is the remaining work. Connectors and document+extraction, if adopted, stay in the importer companion so untrusted input still+crosses the sanitizer boundary and the server remains text-first.++### Current plan after the 2026-10-18 pass++This is the short list that remains after removing work already shipped.+Each item needs its own design and approval before implementation.++1. **P1 - Workstream operations view.** Add a read-only `/web` view over the+ managed-workstream ledger: active, waiting for input, ready to review,+ finished, and stale; show the latest run, handoff, and bounded evidence.+ Store only a minimal human review marker if the derived states are not+ enough. This borrows claude-mem's useful human visibility without turning+ ai-memory into a kanban system or cloud collaboration product.+2. **P1 - Protect interactive recall from background embedding.** Build a+ contention benchmark first. If administrative backfill can delay query+ embedding, give foreground queries reserved capacity or separate gates.+ Preserve the current fail-open behavior: lexical/entity/graph retrieval+ must still answer when embedding or reranking is saturated.+3. **P2 - Structured query filters.** Add bounded, optional filters for path+ prefix, page kind, tier, tag, author, and valid time before ranking. Reuse+ `ScopeResolver`; partial or missing scopes continue to fail closed. This is+ the remaining half of R8 and the most concrete retrieval ergonomics gap.+4. **P2 - Per-operation LLM profiles, measure first.** Prototype separate+ provider/model choices for consolidation, reranking, and auto-improvement,+ plus a cheap structured decision tier. Measure cost, latency, call count,+ and quality before adding public configuration or changing defaults.+5. **P3 - Optional document importer.** If demand is real, extend the+ importer companion to extract PDF/Office/media transcripts into attributed,+ sanitized markdown. Keep opaque files and native multimodal embeddings out+ of the server core.+6. **P3 - End-to-end QA evaluation.** Add a reproducible answerer/judge path+ beside hit@5 and report accuracy, p50/p95 latency, context tokens, exact+ models, and run receipts. Do not claim a cross-tool ranking without a+ shared harness.**Deliberately not recommended:** joining the memory-OS camp (agentself-editing its memory is token-expensive, and Letta's own successor nowkeeps memory as git-backed markdown files);adopting a graph database (bi-temporal-lite on SQLite covers the useful-part); spatial/metaphor architectures (mempalace); publishing any-benchmark number before R2 exists; chasing agentmemory's tool-count+part); spatial/metaphor architectures (mempalace); presenting unlike+benchmarks as a ranking; chasing agentmemory's tool-count(the May conclusion stands: a sharp small surface is the advantage).## 6. Sources@@ -956,7 +954,8 @@ benchmark number before R2 exists; chasing agentmemory's tool-count- LiquidLM: liquidlm.com (+ /about, /docs, /docs/assistants, /docs/mcp: Vaults,MCP tool surface, per-harness support with test dates), github.com/liquidlm(org: an archived empty `cli` repo + a homebrew tap, no engine source), and the- npm registry metadata for `@liquidlm/cli` (MIT, v0.1.5 published 2026-09-17).+ npm registry metadata for `@liquidlm/cli` (MIT; stable v0.1.6 as of+ 2026-10-18, with v0.2.0-alpha.1 as a prerelease).Solo maker (Carlos Souza), closed-source cloud engine, no benchmarks; thebackend internals, model, and any numbers are undisclosed/vendor framing.Analyzed inline in §3 per the no-standalone-doc convention.@@ -1060,3 +1059,31 @@ benchmark number before R2 exists; chasing agentmemory's tool-count2026-07-14).- Adjacent tools: GitHub metadata for codebase-memory-mcp, qmd,context-mode and beads.+- 2026-10-18 delta pass (official sources only):+ - GitHub repository and release metadata for every row in the popularity+ table, plus package registries for `memu-cli` and `@liquidlm/cli`.+ - claude-mem [v14.0.0 release notes](https://github.com/thedotmack/claude-mem/releases/tag/v14.0.0)+ (`My Work`, TeamBrain) and v14.0.1.+ - MemPalace [v3.11.0 release notes](https://github.com/MemPalace/mempalace/releases/tag/v3.11.0)+ and the subsequent default-branch changelog.+ - Hindsight default-branch changelog and+ [mental-model documentation](https://github.com/vectorize-io/hindsight/blob/main/skills/hindsight-docs/references/developer/api/mental-models.md).+ - OpenViking default-branch release notes and docs for multimodal query,+ scoped downloads, Redis scheduling, accounts, privacy, and VikingBot.+ - Cognee [v1.6.5 release notes](https://github.com/topoteretes/cognee/releases/tag/v1.6.5).+ - Graphiti [v0.31.0](https://github.com/getzep/graphiti/releases/tag/v0.31.0)+ and v0.31.1 release notes.+ - Letta V1 README/branch list and Letta Code v0.34.14-v0.34.16 release+ notes. The branch list is the basis for correcting “GitHub-archived” to+ “retired source on the `archive` branch.”+ - EverOS [v1.4.2 release notes](https://github.com/EverMind-AI/EverOS/releases/tag/v1.4.2)+ and changelog.+ - Honcho [changelog](https://github.com/plastic-labs/honcho/blob/main/CHANGELOG.md)+ through v3.3.0.+ - basic-memory default-branch changelog for the unreleased v0.24 section;+ stable release metadata remains v0.23.2.+ - mcp-memory-service [v12.1.1 release notes](https://github.com/doobidoo/mcp-memory-service/releases/tag/v12.1.1)+ and subsequent default-branch commits.+ - Caura's `docs/agent-collaboration-extension.md`, collaboration client+ README, and backend-v3.34.5 release metadata. The docs explicitly call the+ collaboration extension a beta preview and external effects at-least-once.diff --git a/docs/research-basic-memory.md b/docs/research-basic-memory.mdindex 15e5ea88..37b3f4ac 100644--- a/docs/research-basic-memory.md+++ b/docs/research-basic-memory.md@@ -8,6 +8,35 @@> Source project: `basicmachines-co/basic-memory` (Python, MCP-native, markdown-on-disk).> Studied as inspiration *and* as the manual-write-note model we explicitly> diverge from.+>+> **Historical snapshot, refreshed 2026-10-18.** Sections 1-8 describe the+> version originally inspected and preserve file/line evidence from that+> checkout. They are not a claim about current `main`. The latest stable+> release is v0.23.2; the project's default branch carries an unreleased+> v0.24 changelog. Current changes are summarized below.++## 2026-10-18 delta++basic-memory has moved beyond several limitations in the original snapshot.+Its Claude Code plugin now supports session-start briefings, pre-compaction+checkpoints, and opt-in capture guidance. Stable releases also include a local+cross-encoder reranker and a cloud Teams product.++The unreleased v0.24 changelog goes further: optimistic writes with+`expected_checksum`, native OKF v0.2 export/check with generated `index.md`+and `log.md`, cross-project search, structured filters (type, date, category,+metadata, path prefix, and valid time), document extraction with provenance,+temporal qualifiers, a wiki projection, locked notes, more harness+integrations, and real-time cloud collaboration. Those items are on `main`,+not in the latest stable release, so this report does not treat them as+shipped stable behavior.++The useful new comparisons for ai-memory are structured pre-rank filters and+optional document ingestion. ai-memory already ships the underlying ideas in+OKF v0.2, conditional writes, temporal reads, attribution, lifecycle capture,+and git-markdown truth. A document path should remain an importer companion;+live cloud co-editing and a hosted personal knowledge base serve a different+buyer.## 1. Purpose & Scope@@ -109,4 +138,10 @@ For agent handoffs, there is no special handshake; basic-memory treats all MCP c7. **The Markdown source-of-truth tradeoff.** Filesystem latency, checksum recomputation, FTS5 rebuilds, and circuit-breaker retry tracking (`sync_service.py:179-281`) are a lot of infrastructure just to keep a DB consistent with files. For agent memory where the files exist only because LLMs wrote them, this is overhead with no payoff. Keep the DB as primary and only export Markdown on demand.8. **Tool surface is wide (~25 tools).** Agents have to pick among `write_note`/`edit_note`/`move_note`/`delete_note`/`read_note`/`view_note`/`read_content`/`search`/`search_notes`/`build_context`/`recent_activity`/`list_directory`/`list_memory_projects`/`canvas`/`schema_*`. Even with hints, that's a lot of context burned describing tools. Aim for a smaller, more orthogonal set.-**Bottom line**: borrow the *graph + observations + wiki-link* primitives, the *memory:// URI* navigation, the *typed-client + composition-root* layering, the *MCP annotations*, and the *file-as-portable-export* idea, but invert the capture model (ambient, not invoked), add a lifecycle (decay/consolidation/summarization), keep the DB primary, add agent/session provenance, and ship a much narrower MCP tool surface.+**Original bottom line (historical):** borrow the *graph + observations ++wiki-link* primitives, the *memory:// URI* navigation, the *typed-client ++composition-root* layering, the *MCP annotations*, and the+*file-as-portable-export* idea; invert the capture model, add a lifecycle and+agent/session provenance, and keep the tool surface narrow. The old suggestion+to keep the database primary is superseded: ai-memory's established invariant+is markdown-in-git as source of truth with SQLite as a derived index.