FC
FaydabulCloud

Tell your project once. Every AI agent stays up to date.

Faydabul Context syncs Markdown from your repo, versions each publish, and serves search + MCP to Cursor, Claude, ChatGPT, and custom agents.

Start with ContextSee how it works

pnpm add -g @faydabul/cli → faydabul context init → faydabul context sync

Why teams lose context

Docs drift, agents re-read stale files, and nobody agrees which answer is current.

Scattered knowledge

README, Notion, Slack, and old chats disagree. Agents guess.

No version for agents

Production changed Friday; your assistant still cites Tuesday.

Unsafe copy-paste

Secrets in docs and unscoped tokens multiply risk.

Init, sync, connect

1

faydabul context init

Link a context project in faydabul.yml and scan local docs with secret detection.

2

faydabul context sync

Upload only deltas, publish an immutable version, and index chunks for search.

3

Connect agents

Issue read tokens or MCP tools so agents query the same published version with citations.

Built for builders

  • Delta sync from CI or laptop
  • Immutable versions with diff and rollback
  • Full-text search with path, heading, and line citations
  • Scoped read tokens and secret scan (block by default)
  • Collections, share links, and webhooks (rolling out)
  • Hybrid vector search via Qdrant (worker embeddings)

Versions agents can trust

Each sync creates a version you can publish, pin as production, or roll back without rewriting history. Aliases like latest and production keep MCP clients stable.

Security first

Tokens are hashed server-side, revocable instantly, and scoped per project. Document content is treated as untrusted data in MCP responses to reduce prompt injection.

Where teams use Context

Onboarding

New engineers ask agents questions grounded in your runbooks and ADRs.

Support & audit

Pin a version for compliance reviews and share read-only links.

CI documentation

GitHub Actions runs faydabul context sync --ci on every docs change.

Multi-agent products

Several agents share one search index instead of re-uploading files.

Integrations

  • MCP — search_documents, get_document, list_changes
  • HTTP API — /api/context/v1 for custom agents
  • @faydabul/cli CLI — init, sync, watch, search
  • GitHub Actions — example workflow for docs on main
  • Qdrant — vector layer for hybrid search (dev compose)

Example search

Agents receive snippets with stable source URLs.

$ faydabul context search "rollback deploy"
→ docs/runbooks/deploy.md § Rollback (lines 42-58)
→ v12 · score 0.91 · https://…/documents/…

Ship docs once. Let every agent catch up.

Create an account, link a repo folder, and run your first sync in minutes.

Create a free account

Try the first project from the panel

During early access, you can create an account and review the deploy and AI Gateway flow.

Create a free accountSee pricing