Skip to content

Open-source SEO, GEO, and marketing skills for AI agents.

chore: release notfair 0.9.14

Yuting Zhong008c18211d8e
diff --git a/notfair/CHANGELOG.md b/notfair/CHANGELOG.md
index 1e90fa08..3d4b81fd 100644
--- a/notfair/CHANGELOG.md
+++ b/notfair/CHANGELOG.md
@@ -1,5 +1,13 @@
# NotFair
+## 0.9.14 — 2026-07-21
+
+**Goal chat now reflects Codex’s real model defaults and reasoning controls.** The model menu no longer duplicates the configured default, and each provider-supported reasoning effort can be selected per goal and carried safely through the chat API to the Codex CLI.
+
+**Every check exposes the prompt that started it.** The trigger appears as a normal user-style message, expands into fully rendered Markdown, and stays contained within the transcript so long activity labels cannot make the whole workspace scroll.
+
+**Updates download before you ask to apply them.** When a release is available, NotFair downloads it automatically; the visible update button then restarts the managed server and applies the release immediately instead of presenting a second restart step.
+
## 0.9.13 — 2026-07-21
**Adding a server now opens the connector catalog immediately.** The Connections page and onboarding’s **More tools** tile skip the extra menu and go straight to **Browse connectors**, where **Add custom connector** is always the final option—even when every trusted connector is already connected.
diff --git a/notfair/README.md b/notfair/README.md
index 681a755b..dcf6f76e 100644
--- a/notfair/README.md
+++ b/notfair/README.md
@@ -12,7 +12,7 @@ Open source. Runs entirely on your machine. Bring your own LLM credentials (via
- **Fully autonomous, visibly so.** No approval inbox — agents are instructed to act inside the spend envelope you set, with observation-window discipline, a pause button for scheduled work, and their recorded moves visible in the app. These are workflow controls for a trusted harness, not a sandbox against arbitrary local commands.
- **Code changes through pull requests.** Attach a GitHub-backed codebase and an agent can work in its own branch, commit and push the change, then register the pull request with NotFair. The app tracks the PR, while its goal protocol instructs the agent to leave merge decisions to you and keep one code mutation in flight at a time. Those rules are prompt-level guardrails, not mechanically enforced permissions.
- **Shared context + private memory.** `PROJECT.md` is the workspace brief every agent carries (any agent can update it via `set_shared_context`); each agent also keeps its own learnings ledger and workspace files. All connected MCPs are shared by every agent.
-- **One screen per goal.** The conversation and the loop's state live together: chat on the left (where the goal is defined and steered), a status rail on the right — the progress chart, the plan, every check with its own follow-up chat, open actions with review dates, and the agent's memory. No tabs, no thread management, nothing else to learn.
+- **One screen per goal.** The conversation and the loop's state live together: chat on the left (where the goal is defined and steered), a status rail on the right — the progress chart, the plan, every check with its own follow-up chat, open actions with review dates, and the agent's memory. Opening a check also exposes the fully rendered Markdown prompt that triggered it, so the original instruction and follow-up conversation stay together. No tabs, no thread management, nothing else to learn.
- **Progress you can see.** A time-true chart with the target line, every agent action as a marker on the moment it happened (hover: what it did, what it predicted, what actually happened), observation windows shaded, and history backfilled at setup from the platform's own date-segmented stats — context from day one. Maintain goals get a streak ("held at target for 12 days") with a per-check strip; the workspace index shows a mini sparkline + 7-day delta per goal.
- **Project-scoped MCP connections** — one-click PKCE OAuth brings third-party tools (Google Ads via NotFair's hosted MCP) into the agents' toolbox. Connection records are stored in local SQLite and wired into the chosen harness automatically; Codex receives bearer tokens through per-server environment variables at process launch.
@@ -22,7 +22,7 @@ At onboarding you pick which local AI coding agent runs the work:
| Harness | Status | Notes |
|---|---|---|
-| **Codex** | Recommended | Uses your existing `codex` login. Per-server env-var bearers. The sidebar reports the authenticated account's plan and usage, and model selectors show the configured model name. The adapter launches Codex with `--dangerously-bypass-approvals-and-sandbox`; that process inherits the parent environment and can access files available to your local user. Treat goal agents as trusted local automation. |
+| **Codex** | Recommended | Uses your existing `codex` login. Per-server env-var bearers. The sidebar reports the authenticated account's plan and usage; model selectors use Codex's configured/provider metadata, show the default once, and expose each model's supported reasoning efforts per goal. The adapter launches Codex with `--dangerously-bypass-approvals-and-sandbox`; that process inherits the parent environment and can access files available to your local user. Treat goal agents as trusted local automation. |
| **Claude Code** | Supported | Uses your existing `claude` login. Per-agent `.mcp.json` for isolation. |
Different projects can run on different harnesses; the choice persists on the project row.
@@ -77,7 +77,7 @@ notfair --help
Options on `start`: `--port <n>` (default 3327), `--no-open`, `--foreground`, `--data-dir <path>`.
Options on `doctor`: `--port <n>`, `--data-dir <path>`.
-**Updating.** `notfair update` checks npm, installs the newer version globally, and restarts whatever is running — through launchd when autostart is enabled (rewriting the LaunchAgent to the fresh install), or by stopping and relaunching the background daemon. The app has the same flow built in: the sidebar footer shows the running version, an "✨ vX.Y.Z available" button appears when npm has a newer release, and clicking it installs the update and then offers **Restart now** — the server restarts itself (launchd- and daemon-managed runs) and the page reloads on the new version. Foreground/dev runs are never killed from the app; they get a "restart in your terminal" note instead.
+**Updating.** `notfair update` checks npm, installs the newer version globally, and restarts whatever is running — through launchd when autostart is enabled (rewriting the LaunchAgent to the fresh install), or by stopping and relaunching the background daemon. The app has the same flow built in: when npm reports a newer release, the sidebar downloads it automatically and then shows **Update to vX.Y.Z**. Clicking that button restarts a launchd- or daemon-managed server, applies the release, and reloads the page immediately; there is no second restart step. Foreground/dev runs are never killed from the app; they get a "restart in your terminal" note instead.
When autostart is enabled, launchd owns the server: `notfair start` delegates to it (never spawns a competing copy), `notfair stop` stops it until your next login, and `notfair autostart disable` removes it entirely. The LaunchAgent captures your shell's PATH so the `claude` / `codex` binaries stay reachable at login, and `notfair start` self-heals the entry if an upgrade moved the package on disk. Global install is recommended for autostart — the npx cache can be cleared at any time.
diff --git a/notfair/package.json b/notfair/package.json
index e6c00c6d..30529098 100644
--- a/notfair/package.json
+++ b/notfair/package.json
@@ -1,6 +1,6 @@
{
"name": "notfair",
- "version": "0.9.13",
+ "version": "0.9.14",
"description": "Goal-driven, loop-powered marketing agents that crush your business goals 24/7 — on your own machine. Say 'cut CAC to $30' and a dedicated agent turns it into a server-verified metric, then moves the number check after check while you sleep. Rides your existing Claude Code or Codex login.",
"license": "MIT",
"bin": {