## Summary
- Switch the docs Eve agent from the AI Gateway `openai/gpt-5.4-mini`
string to an Inception Labs Mercury 2 OpenAI-compatible chat model.
- Keep tool calling on the chat-completions path and pass Mercury's
`reasoning_effort=medium` through the AI SDK OpenAI adapter.
- Add `DOCS_AGENT_MODEL_FLOW` so the same agent can run either `mercury`
or the old AI Gateway flow for eval comparisons.
- Add docs-agent eve evals covering grounded docs answers, docs
retrieval, citations, and account-specific support refusal.
- Replace the docs-agent retriever with an in-process BM25-style lexical
ranker that returns bounded full content for the top results, so Mercury
gets rich context in one fast tool call instead of a serial
`search_docs` → `read_doc` round trip.
- Precompute BM25 term counts/document frequencies into the generated
`agent/lib/docs-index.ts` snapshot at build time, removing deployed
cold-start corpus construction while keeping retrieval in-process.
- Add opt-in search perf logging (`DOCS_AGENT_SEARCH_PERF_LOG=1`,
optional `DOCS_AGENT_SEARCH_LOG_QUERY=1`) with timings for tokenization,
corpus load/cache, ranking, hydration, total duration, corpus source,
and top URLs.
- Add `eval:agent` and `eval:agent:flows` scripts; `eval:agent:flows`
can run local model-flow comparisons or remote target comparisons via
`DOCS_AGENT_EVAL_TARGETS`.
- Add `INCEPTION_API_KEY` / optional Mercury and gateway model knobs to
`docs/.env.example`, and move `@ai-sdk/openai` to runtime dependencies
for the agent import.
## Notes
- This is intentionally an experiment to see how Mercury's diffusion
model behaves with Eve tool calling (`search_docs` and `read_doc`).
- Preview/runtime environments need `INCEPTION_API_KEY`;
`INCEPTION_MODEL` and `INCEPTION_BASE_URL` are optional overrides.
- The custom fetch prevents accidentally falling back to
`OPENAI_API_KEY` against Inception's endpoint.
- The docs search is lexical/in-memory, not vector search. The slow path
was mostly serial model/tool round trips and cold index construction,
not embedding lookup.
- The generated BM25 snapshot is process-local once loaded: warm for the
lifetime of the running Node/Vercel function instance, and reset on cold
starts, redeploys, or process restarts. The expensive term-count corpus
is now built at docs build time.
- Perf logs omit raw user queries by default; set
`DOCS_AGENT_SEARCH_LOG_QUERY=1` only when you explicitly want raw
query/term logging.
- Local A/B-style eval run:
```bash
DOCS_AGENT_EVAL_FLOWS=gateway,mercury bun run eval:agent:flows --
--strict
```
- Live target comparison:
```bash
DOCS_AGENT_EVAL_TARGETS=baseline=https://<prod>,mercury=https://<preview>
bun run eval:agent:flows -- --strict
```
## Tests
- `bunx eslint scripts/build-agent-index.ts agent/lib/docs.ts
agent/tools/search_docs.ts`
- `bun scripts/build-agent-index.ts` (wrote 133 pages + 1000 toolkits +
1139 BM25 rows)
- `DOCS_AGENT_SEARCH_PERF_LOG=1 EVE_FORCE_BUNDLE=1 bun -e "const
tool=(await import('./agent/tools/search_docs.ts?log=' +
Date.now())).default; await tool.execute({query:'create a session with
github tools'}); await tool.execute({query:'auth config connected
account'});"` (logs cold and warm timing JSON)
- `EVE_FORCE_BUNDLE=1 bun -e "const tool=(await
import('./agent/tools/search_docs.ts?bundle=' + Date.now())).default;
const started=performance.now(); const r=await
tool.execute({query:'create a session with github tools'});
console.log(r.retrieval, r.results[0].url, r.results[0].content.length,
Math.round(performance.now()-started)+'ms');"` (precomputed bundle path,
~12ms)
- `bun -e "const tool=(await import('./agent/tools/search_docs.ts?live='
+ Date.now())).default; const started=performance.now(); const r=await
tool.execute({query:'create a session with github tools'});
console.log(r.retrieval, r.results[0].url, r.results[0].content.length,
Math.round(performance.now()-started)+'ms');"` (live-content path,
~34ms)
- `EVE_FORCE_BUNDLE=1 bun -e "const tool=(await
import('./agent/tools/search_docs.ts')).default; await
tool.execute({query:'create a session with github tools'}); const
started=performance.now(); const r=await tool.execute({query:'auth
config connected account'}); console.log(r.results[0].url,
r.results[0].content.length,
Math.round(performance.now()-started)+'ms');"` (warm path ~2ms)
-
`PATH=/Users/cryogenicplanet/.vite-plus/js_runtime/node/24.15.0/bin:$PATH
./node_modules/.bin/eve info --json` (reports `status: ready`, `model:
inception/mercury-2`, `errors: 0`)
- `DOCS_AGENT_MODEL_FLOW=gateway
PATH=/Users/cryogenicplanet/.vite-plus/js_runtime/node/24.15.0/bin:$PATH
./node_modules/.bin/eve info --json` (reports `status: ready`, `model:
openai/gpt-5.4-mini`, `errors: 0`)
-
`PATH=/Users/cryogenicplanet/.vite-plus/js_runtime/node/24.15.0/bin:$PATH
./node_modules/.bin/eve eval --list`
-
`PATH=/Users/cryogenicplanet/.vite-plus/js_runtime/node/24.15.0/bin:$PATH
bun scripts/eval-agent-flows.ts --list`
- `bun test tests/static/` (16 passed)
- `bun run types:check` currently fails on existing docs type-generation
errors in `app/(home)/docs/changelog/[...slug]/page.tsx`,
`app/(home)/examples/[[...slug]]/page.tsx`,
`app/(home)/toolkits/[[...slug]]/page.tsx`,
`app/llms.mdx/[[...slug]]/route.ts`, `lib/search-index.ts`, and
`lib/source.ts`; no new eval or `docs/agent/agent.ts` errors were
reported.
## Not run
- Real live model evals, because this local environment does not have
`INCEPTION_API_KEY` or AI Gateway credentials.
## Latest update
- Added default eager docs retrieval in the Eve HTTP channel: the server
runs the same BM25 search on the user's message before the first model
step and injects the results as one-turn context.
- Kept `search_docs` and `read_doc` available so Mercury can still
search/read more when the eager context is weak, ambiguous, or missing.
- Added `DOCS_AGENT_EAGER_SEARCH=0` as an escape hatch and labeled perf
logs with `invocation: "eager_context" | "tool"`.
- Updated the loading copy from “Searching the docs…” to “Thinking with
the docs…” so UI latency is not attributed solely to the search call.
## Latest tests
- `bun run lint -- agent/channels/eve.ts agent/tools/search_docs.ts
agent/lib/docs-search.ts components/eve-chat.tsx
evals/docs-agent/grounded-answers.eval.ts`
-
`PATH=/Users/cryogenicplanet/.vite-plus/js_runtime/node/24.15.0/bin:$PATH
node_modules/eve/bin/eve.js info --json` (reports `status: ready`,
`errors: 0`)
- `DOCS_AGENT_SEARCH_PERF_LOG=1 EVE_FORCE_BUNDLE=1 bun -e "import {
searchDocs } from './agent/lib/docs-search'; const r = searchDocs('How
do I create a session in Composio? Keep it brief.', { invocation:
'eager_context' }); console.log(JSON.stringify({count:r.results.length,
top:r.results[0]?.url, content: !!r.results[0]?.content}, null, 2));"`
- `DOCS_AGENT_SEARCH_PERF_LOG=1 bun -e "import { searchDocs } from
'./agent/lib/docs-search'; searchDocs('How do I create a session in
Composio? Keep it brief.', { invocation: 'eager_context' });
searchDocs('How do I create a session in Composio? Keep it brief.', {
invocation: 'tool' });"`
- `bun run types:check` still fails only on the pre-existing docs
type-generation issues listed above.
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Integration branch for the next docs release: a **sessions-first
documentation rewrite** — new and rewritten guides, example pages,
interactive components, and docs tooling — plus the supporting SDK
changes that the new docs describe.
The bulk of this PR is docs (~24k lines across ~150 commits); the SDK
changes (~5k lines) back the new guides.
## Documentation (the bulk)
- **Sessions-first restructure** — reorganized navigation and section
structure (incl. the "Sandbox (prev workbench)" section), with
v3-reorganization redirects so old URLs keep resolving.
- **Rewritten core guides** — quickstart, configuring sessions, triggers
(creating + subscribing to events), proxy-execute, toolkits
enable/disable, and common FAQ, rewritten in the house voice.
- **New example pages** — local-sandbox PR reviewer, daily standup bot,
and slack bot, with runnable build-ups.
- **New interactive components & diagrams** — triggers flow animation,
manage-connections visual, connection-refresh visual, and the
terminal-kit components.
- **Docs tooling** — a docs-graph link-graph connectivity checker,
search reprioritization (deprioritize legacy pages), and SDK-reference
regeneration.
## Supporting SDK changes
**`@composio/core` → 0.13.0 (minor)**
- `composio.sessions.create()` as the first-class sessions API
(`composio.create()` kept as an alias).
- **MCP is opt-in:** default `create()` / `use()` return native-tool
sessions (`SessionWithoutMcp`); pass `{ mcp: true }` to surface
`session.mcp`. _Migration: read `session.mcp` only after creating with
`{ mcp: true }`._
- `session.sandbox` is the canonical resolved config;
`session.workbench` kept as a deprecated alias. `sandbox` is the
preferred session-config key (`workbench` still accepted).
- `connectedAccounts.updateAcl()` graduated from experimental (alias
kept).
- `triggers.parse()` (parse + optionally verify an incoming webhook) and
`triggers.setWebhookSubscription()`.
**`@composio/experimental` → minor** — local-workbench helpers moved
onto the `@composio/experimental/workbench` subpath (out of
`@composio/core/experimental`), keeping the ~14 KB embedded Python
helper out of core. Plus the experimental Pi provider.
**`@composio/slim` → minor.**
**Python → 0.17.0** — mirrors the TS surface: `composio.sessions` mount
(`tool_router` deprecated), `triggers.parse()` /
`set_webhook_subscription()`, the `sandbox` config key, and
`connected_accounts.update_acl()`.
## Review response (#3664)
Addressed the `@composio/core` review:
- **Security:** `triggers.parse()` no longer fails open — a
present-but-empty `verifySecret` (e.g. unset `COMPOSIO_WEBHOOK_SECRET`)
now throws instead of silently skipping verification; omitting it stays
an explicit opt-out (both SDKs).
- Removed snake_case leakage from `transformWebhookSubscription` (+ the
index signature that allowed it).
- **Removed** the TS-only `connectedAccounts.link()` toolkit
auto-resolve (shipped with cancellability / orphaned-auth-config bugs
and was effectively undocumented; to be reintroduced properly later).
- Unified Python error types on `ValidationError`; added `mcp=True`
Python tests; fixed runtime-portability + error-type test assertions.
- Polished deprecation messages; fixed the backwards `/experimental`
`@deprecated` note and the `SessionWithMcp` JSDoc.
## Testing
- **TS:** `@composio/core` + `@composio/experimental` typecheck pass;
vitest green for the touched suites.
- **Python:** `test_tool_router.py` + `test_triggers.py` pass (161
tests).
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Kshitij Jhunjhunwala <kj@composio.dev>
Co-authored-by: Malay Vasa <malayvasa@gmail.com>
Co-authored-by: Sarah Simionescu <sarah@composio.dev>
Co-authored-by: Kshitij Jhunjhunwala <113939507+KJ-11@users.noreply.github.com>
This PR:
- consolidates and bumps TypeScript/npm dependencies across the
monorepo, docs, examples, and e2e fixtures — no runtime behavior changes
- **cleanup:** remove the unused `ansis` dependency from `@composio/cli`
(`picocolors` is the actual color lib), drop the dead `uuid` catalog
entry, catalog `dotenv` + `@types/bun` and repoint drifting examples/e2e
onto them, and replace `chalk` with `picocolors` in `@composio/core`
(smaller, ESM, already used by the CLI)
- **TypeScript 6:** bump `typescript` `5.9 → 6.0.3` everywhere (catalog,
CLI test fixtures, docs); drop vestigial `declaration`/`outDir` from the
provider `tsconfig.json`s to fix the TS 6 `rootDir` regression
(`TS6059`); add `ignoreDeprecations: "6.0"` in docs for the `baseUrl`
deprecation
- **toolchain:** `tsdown 0.18 → 0.22.3`, `vitest` + `@vitest/ui →
4.1.9`, `publint → 0.3.21`, `wrangler → 4.101.0` (each the latest
version within the 3-day `minimumReleaseAge` gate)
- **Effect + hono:** `effect 3.21.3`, `@effect/cli 0.75.2`,
`@effect/platform 0.96.1`, `platform-bun 0.90.0`, `platform-node-shared
0.60.0`, `language-service 0.86.2`, `@effect/vitest 0.29.0`, `hono
4.12.25`; pin the Effect peer cohort
(`printer`/`printer-ansi`/`typeclass`/`rpc`/`sql`/`cluster`/`experimental`/`workflow`)
as `@composio/cli` devDeps so the auto-installed peers resolve
coherently, and bump `@cloudflare/workers-types` to satisfy `wrangler`'s
peer
- changesets: `@composio/cli` patch (ansis removal) and `@composio/core`
patch (chalk → picocolors)
- verified: `build`, `typecheck` (tsgo + real `tsc` 6.0.3), and local
tests (excluding e2e) all pass
## Summary
Reworks the docs welcome page (`/docs`) into a two-column hero with an
animated agent chat mock + tool overlay, plus four restructured homepage
sections.
### Hero (`DocsHeroV2`)
- Left: headline (`Start building with Composio.`), short sub,
brand-blue chip column (Quickstart, API reference, Toolkits).
- Right (lg+ only): mock chat panel with the v1 hero's per-agent skin
(`Claude Cowork` / `ChatGPT Codex` / `Your AI Agent`), step-by-step
timeline-driven animation (user prompt → thinking dots → tool-trail
items → assistant response), and a floating "active tool" card layered
over the chat's left edge.
- L-shape SVG connector between chat and overlay: muted grey resting
state + brand-blue glow + bright line that re-draw via `pathLength` on
every tool change.
- Mock chat hidden below lg.
### Homepage sections
- **`HomeFeatures`** — `Everything you need to ship production agents.`
2×2 grid with bespoke visuals per pillar: 8×2 toolkit grid with `+1K`
cell, sessions panel (`USER_ID` + connected apps), triggers feed (`LIVE`
indicator + recent events), and a workbench mock with animated 4×2 CPU
lights. `auto-rows-fr` on `sm:` so cards equalize on tablet/desktop but
breathe on mobile.
- **`HomeSurfaces`** — `Three ways to use Composio.` bento: Developer
card spans full width (`md:col-span-3`) and shows 8 provider logos
(Anthropic, OpenAI, Vercel AI, Google, LangChain, CrewAI, LlamaIndex,
Mastra). Row 2: For You (with Claude / Codex / Cursor / Windsurf /
OpenClaw client logos), CLI (with `curl ... | bash` install snippet and
functional copy button), Agents.
- **`HomeResources`** — `Reach for the rest.` six-card footer row with
Quickstart, API reference, Cookbooks, Changelog, Toolkits,
Troubleshooting.
### Other
- Ports the landing-page hero into `components/landing-hero/` (kept
under the original `DocsHero` for easy revert; commented out, V2 wired
in).
- Theme-aware `--hero-*` CSS vars for chat-skin colors so the chat reads
in both light and dark.
- Removes the old `Works with your stack` provider grid (the providers
now live inside the Developer surface card).
## Test plan
- [ ] Verify hero animation cycles through 3 examples on desktop
- [ ] Verify L-shape connector pulses in brand blue on every tool change
- [ ] Verify chat mock is hidden on mobile and left column reads on its
own
- [ ] Verify dark mode (chat skins, connector, card surfaces, code
blocks)
- [ ] Verify CLI install copy button copies and toggles the icon to ✓
- [ ] Verify all chip / card hrefs resolve
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
## Summary
- Re-apply the Algolia docs search migration after the previous PR was
merged and reverted.
- Keep the existing Fumadocs search UI while using Algolia API when
search keys are configured, with `/api/search` fallback for local
development and tests.
- Add a first-party Algolia index builder/sync script that creates
section-sized docs records from MDX/OpenAPI/toolkit/changelog content,
configures index relevance settings, and replaces index objects without
relying on Algolia Crawler.
- Add Algolia Insights view/click events and a terminal search relevance
test script.
- Clean search breadcrumbs so results show labels like `Toolkit` and
`Cookbook` instead of duplicated `toolkits > Gmail` formatting.
## Tests
- `cd docs && bun run types:check`
- `cd docs && bun run sync:search --dry-run`
- `cd docs && bunx eslint components/custom-search-dialog.tsx
lib/search-index.ts scripts/sync-algolia-search.ts
scripts/test-algolia-search.ts`
## Notes
- Live Algolia sync requires `ALGOLIA_ADMIN_API_KEY`.
- Live search relevance tests require `ALGOLIA_SEARCH_API_KEY` or
`NEXT_PUBLIC_ALGOLIA_SEARCH_API_KEY`.
## Summary
- Switch docs search dialog to Algolia when public Algolia env vars are
configured, with local `/api/search` fallback for development/tests
- Add a shared docs search index builder plus `bun run sync:search` to
publish records to Algolia
- Add a GitHub Actions workflow to dry-run and sync the Algolia index on
`next` docs changes when secrets are configured
- Document required Algolia env vars and sync command
## Tests
- `cd docs && bun run types:check`
- `cd docs && bun run sync:search --dry-run`
- `cd docs && bun test tests/static/`
- `cd docs && bunx eslint components/custom-search-dialog.tsx
lib/search-index.ts scripts/sync-algolia-search.ts`
- `cd docs && bun run build`
## Notes
- `bun run lint` still fails on existing unrelated repo-wide lint errors
in files outside this change (for example `app/global-error.tsx`,
`components/ask-ai-button.tsx`, `components/version-selector.tsx`). The
changed search files pass targeted ESLint.
- Production search requires `NEXT_PUBLIC_ALGOLIA_APP_ID`,
`NEXT_PUBLIC_ALGOLIA_SEARCH_API_KEY`, and optionally
`NEXT_PUBLIC_ALGOLIA_INDEX_NAME` (default `composio_docs`). Index
syncing requires `ALGOLIA_APP_ID`/`ALGOLIA_ADMIN_API_KEY` secrets.
New page at /docs/auth-configuration/shared-connections covers the
account_type and per-user ACL surface added to the Connected Accounts
SDK in 0.9.1:
- When to use SHARED vs PRIVATE (table)
- Creating a SHARED connection via link() with an initial ACL
- The 4-step deny-wins resolution rule + common ACL patterns
- updateAcl() with PATCH semantics + empty-array footgun warning
- Using a shared connection from a tool-router session (pinning) or
via direct execute with connectedAccountId
- Inspecting account_type + aclConfigForShared on get() responses
- Error handling table covering ComposioAclOnlyForSharedError,
ComposioSharedAccessDeniedError, ComposioSharedConnectionNotAccessibleError
Other changes:
- Bump docs/package.json @composio/core dependency from ^0.9.0 to ^0.9.1
so twoslash type-checks against the version that exposes the new
fields.
- Wire the new page into the Authentication sidebar between
connected-accounts and migrating-initiate-to-link.
Verified: bun run build clean (1411 pages including the new one).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Bumps `@composio/core` devDependency to ^0.8.1 (the release that ships
`workbench.sandboxSize` and `connectedAccounts.link({ allowMultiple })`)
so the new snippets in this branch type-check end-to-end. Removes the
`// @noErrors` directives that were only there because the SDK predated
those APIs.
While bumping, also fixes the toolkits pagination snippet in
fetching-tools-and-toolkits.mdx for the 0.6 -> 0.8 rename of
`nextCursor` -> `cursor` on `session.toolkits()`, and adds explicit
`enable: true` to the `workbench` examples that set `sandboxSize`
(the inferred config type requires `enable`).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Co-authored-by: abir <abir@composio.dev>
SDK 0.6.10 includes the ToolRouterAuthorizeFn alias fix (#3167),
so all code blocks now type-check cleanly without suppression.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Zen Agent <zen@composio.dev>
- Bump @composio/core to ^0.6.9 for multiAccount type support
- Add // ---cut--- preamble to changelog TypeScript code blocks
- Add // @errors: 2353 for session.authorize alias (pending SDK type fix in #3167)
- Add missing session variable to Python code blocks
- Add Python snake_case names to config options table
- Fix changelog intro sentence fragment
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Zen Agent <zen@composio.dev>
- Bump @composio/core to ^0.6.9 so connectedAccounts.update() type-checks
- Add import + cut directive to changelog TS code block (fixes twoslash error 2304)
- Update .gitleaksignore with correct commit hashes after rebase
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace the verbose auth config / connected accounts pattern with the
modern session API (composio.create, session.tools, session.authorize,
session.toolkits). Switch to OpenAI Agents SDK for the agentic loop.
Also install @ai-sdk/react to fix twoslash build error in chat-app tutorial.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add jsx: react-jsx compiler option to twoslash config (fixes React UMD global errors)
- Install @ai-sdk/react for useChat type resolution
- Exclude examples/ from tsconfig (twoslash handles type-checking via <include>)
- Allow expected error 2307 for local component import in page-with-tools.tsx
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Rename experimental_createMCPClient to createMCPClient (graduated in @ai-sdk/mcp v1.0)
- Fix multi-turn message history to include tool calls via result.response.messages
- Add missing client.close() cleanup in MCP snippet
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Rename experimental_createMCPClient to createMCPClient (graduated in @ai-sdk/mcp v1.0)
- Fix multi-turn message history to include tool calls via result.response.messages
- Add missing client.close() cleanup in MCP snippet
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add loading="lazy" and decoding="async" to toolkit icon images
so the browser only fetches them as they scroll into view instead
of requesting all ~855 icons at once on initial load.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add a new docs page explaining the organization/project hierarchy,
project management via API, and project settings. Also set up mermaid
diagram rendering for the docs site (remark plugin + client component).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace WebhookEventTypes enum and ConnectionExpiredEventSchema with
plain string literals since they aren't available in the published SDK yet.
Update @composio/core from 0.5.5 to 0.6.3.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add @vercel/analytics for client-side page view tracking and
server-side 404 logging to Datadog from the not-found page.
This enables monitoring of broken links via Datadog log queries.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The @composio/vercel tool types are incompatible with ai@6.0.49.
Added @noErrors directive to Vercel AI examples - the code is
functionally correct, just has type mismatches between packages.
Also simplified message content format back to strings.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Remove Direct MCP usage accordion (complex API, not recommended)
- Keep Native Tools and MCP main sections with session-based approach
- Keep Direct tool usage accordion for single tool use cases
- Update @composio/* packages from 0.4.0 to 0.5.5
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Rahul Tarak