Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors
Rust
docs: polish PR 21 prose with humanizer
Aldo Junioraabe8a324526
diff --git a/CHANGELOG.md b/CHANGELOG.mdindex 8f04f769..57d21b8d 100644--- a/CHANGELOG.md+++ b/CHANGELOG.md@@ -23,14 +23,14 @@ version in [`docs/changelog/`](docs/changelog/):credentials. The source-built Docker image now copies `CHANGELOG.md` for this(#NNN).-- Project pages now show one section at a time in accessible tabs: Overview,+- Organized project pages into accessible tabs showing one section at a time: Overview,Agents, Token usage, Sessions and Knowledge. The selected tab is kept in theURL (`?tab=agents`), which keeps reload, shared links and browser history insync, and the tabs support Left, Right, Home and End. Existing`#agents`-style section bookmarks still open the matching tab (#NNN).-- The browser inbox composer now sets To project to the source project by- default and lists only harnesses with captured sessions in the destination+- Updated the browser inbox composer to default To project to the source+ project and list only harnesses with captured sessions in the destinationworkspace and project. It resets a stale recipient when a project changes,and it disables sending, with an explanation, for a project that has noregistered agent (#NNN).diff --git a/docs/features.md b/docs/features.mdindex 7f3eeb7e..1781f648 100644--- a/docs/features.md+++ b/docs/features.md@@ -1,7 +1,7 @@# Features and activation-Use this guide to choose a feature, enable its prerequisites, and check the-result. It describes shipped behavior; the linked guides contain full options.+This guide covers prerequisites, activation and verification for shipped+features. Follow the linked guides for the full options.Configuration examples belong in `<data_dir>/config.toml` unless a section names`.ai-memory.toml`. Restart the server after changing server configuration.@@ -25,8 +25,8 @@ server container does not update clients or hooks on a separate machine.## Capture and agent integration-Purpose: capture bounded, sanitized lifecycle observations from an agent.-Prerequisite: a reachable server and a supported harness installed locally.+Lifecycle hooks capture bounded, sanitized observations from an agent. They+require a reachable server and a supported harness installed locally.The managed launcher can wire supported integrations on first launch:@@ -53,9 +53,9 @@ registered sessions and captured-event counts should grow. If they do not, use## Token accounting and investigation-Purpose: inspect usage for individual model responses and totals per session-and agent type. Enable: update the native Claude Code or Codex client and-reinstall its hooks as above. There is no separate dashboard token switch.+The dashboard shows usage for individual model responses and totals per session+and agent type. Update the native Claude Code or Codex client and reinstall its+hooks as above to enable capture. There is no separate dashboard token switch.The server records reports as they arrive; the session need not end.Verify: open the project, then the agent, then the session, and inspect "Token usage". The project@@ -69,8 +69,8 @@ they are not confirmed failures or monetary costs. See [dashboard usage](usage.m## Themes and responsive navigation-Enable: choose "Light", "Dark" or "System" in the header. The preference-is saved locally in that browser. No server setting or restart is required.+Choose "Light", "Dark" or "System" in the header. The browser saves your+preference locally. No server setting or restart is required.On small screens, open the menu for projects, inbox, search and documentation.Inside a project, "Overview", "Agents", "Token usage", "Sessions" and@@ -84,9 +84,9 @@ destinations off-screen.## Agent inbox and directed messages-Purpose: queue a request for a project, optionally targeting a harness such-as Codex or Claude Code. Prerequisite: the recipient must have a captured-session in that project to appear in the browser selector.+Queue a request for a project, optionally targeting a harness such as Codex+or Claude Code. The recipient must have a captured session in that project to+appear in the browser selector.Open "Inbox" and choose "From project". "To project" defaults to the sameproject; change it explicitly to send across projects. "To agent" lists only@@ -121,8 +121,8 @@ shared queues, cancellation, sender provenance and untrusted-message handling.## Handoffs and compaction recovery-Purpose: carry useful context into the next session or harness.-Enable: install lifecycle hooks. Handoff capture and recovery follow the+Install lifecycle hooks to carry useful context into the next session or+harness. Handoff capture and recovery follow theconfigured lifecycle policy; no LLM is required for the deterministic path.Use `ai-memory run` for managed cross-harness continuity.@@ -132,8 +132,9 @@ is not evidence of a running process. See [handoffs and recovery](usage.md).## Managed sessions and workstreams-Enable: launch a supported harness with `ai-memory run <harness>` rather-than a direct executable. This is opt-in per launch, not a global mode switch.+Launch a supported harness with `ai-memory run <harness>` to enable managed+sessions for that launch. Direct launches remain unmanaged; this setting does+not change the global mode.Use the documented resume and launch-profile options for account-specificconfiguration. Root can inspect recorded lifecycle metadata in "Workstreams".@@ -164,8 +165,8 @@ and the [configuration reference](ARCHITECTURE.md#configuration-configtoml).## Embeddings and reranking-Purpose: add semantic retrieval and optional candidate reranking.-Enable: configure the embedding provider, model and credentials described in+For semantic retrieval and optional candidate reranking, configure the+embedding provider, model and credentials described in[LLM providers](llm-providers.md) and the [configuration reference](ARCHITECTURE.md#configuration-configtoml).Availability and defaults depend on the provider configuration and build.LLM reranking is separately opt-in through `AI_MEMORY_RERANKER=llm` and requires@@ -177,9 +178,9 @@ vectors are ignored. Ordinary lexical recall remains available without them.## LLM consolidation and provider fallback-Purpose: compile captured evidence into maintained knowledge with an optional-model provider. Enable: configure `llm_provider`, model and the provider's-credential flow from [LLM providers](llm-providers.md). Provider support and+To compile captured evidence into maintained knowledge with an optional model+provider, configure `llm_provider`, the model and the provider's credential flow+from [LLM providers](llm-providers.md). Provider support andcredential types are documented there; do not paste credentials into pages.Without an LLM, capture, lexical recall and deterministic summaries still work.