What is Indexed
What is Indexed
Indexed is a privacy-first CLI tool that makes your team's institutional knowledge — scattered across Jira tickets, Confluence pages, and local documents — discoverable through semantic search. It runs entirely on your machine: documents are parsed, embedded into vectors using a local AI model, and stored in a FAISS index. Nothing leaves your machine. The index is then exposed to AI assistants like Claude Desktop and Cursor via the Model Context Protocol (MCP), so your AI tools can search your company's knowledge base as naturally as they search the web.
Use Cases
Your Jira tickets are a black hole
Your team has two years of Jira tickets — decisions, bug investigations, architecture discussions, deployment runbooks. A new engineer joins and asks "how do we handle auth token rotation?" The answer is in PROJ-1847 from eight months ago, but nobody can find it because Jira search is keyword-based and the ticket title says "SSO refresh logic." Indexed makes all of it semantically searchable — search "auth token rotation" and find that ticket instantly.
Your AI assistant can't see your internal docs
You use Claude Desktop daily, but it has no idea your team wrote a deployment guide, a database migration runbook, or a detailed incident postmortem last quarter. Indexed exposes all of that as MCP tools, so when you ask Claude "how do we roll back a migration?", it pulls the answer from your own docs.
Your Confluence wiki is a graveyard
Hundreds of pages of onboarding guides, architecture decisions, and API documentation — and nobody can find any of it because Confluence search barely works. Indexed brings it back to life with meaning-based search that understands what you're looking for, not just the exact words you type.
How It Works
┌─────────────────┐ ┌──────────────────┐ ┌─────────────────────┐
│ Your Sources │ │ Indexed CLI │ │ Search & MCP │
│ │────▶│ │────▶│ │
│ Local Files │ │ Parse documents │ │ indexed index │
│ Jira Tickets │ │ Chunk text │ │ search "query" │
│ Confluence Pages│ │ Embed locally │ │ │
│ │ │ Store in FAISS │ │ Claude Desktop │
│ │ │ │ │ Cursor / Cline │
└─────────────────┘ └──────────────────┘ └─────────────────────┘
All processing happens
on YOUR machine.- Connect your sources — point Indexed at a folder, a Jira project, or a Confluence space
- Index — documents are parsed, split into chunks, and embedded into vectors using a local model (no API calls, no data sent anywhere)
- Search — query your index from the CLI or let AI assistants search it via MCP
Privacy by Design
Your data never leaves your machine
Indexed runs a local embedding model (all-MiniLM-L6-v2) directly on your machine. The only network calls are to your own Jira/Confluence instances when fetching documents. No data is sent to Indexed, HuggingFace, or any third party. No telemetry. Everything is stored locally at ~/.indexed/.
Read more about privacy and data storage.