Advanced search controls
Set the rules for relevant results. Filters and ranking controls help match each query with the right information.
Query storage buckets and cloud drives.
Power agentic search + RAG at scale.
claude mcp add --transport http captain https://mcp.captain.dev/mcp claude mcp login captain
Building your first RAG pipeline is fun and easy to vibe code. But scaling it and keeping it in sync tends to be where these things fail.
Agentic search often costs 3x less on Captain and delivers better accuracy.
Ditch zero-shot RAG and call the Query API as a tool until you find what you need.
Agentic search often costs 3x less on Captain and delivers better accuracy.
Ditch zero-shot RAG and call the Query API as a tool until you find what you need.
Sync information and self-improve search.
Every layer above ships together, on one index. Nothing to wire and nothing to keep aligned.
Used by Y Combinator, compliance teams, and engineers building agents that need reliable answers.
The accuracy was noticeably better. Setup took 5 minutes, and it scaled to thousands of docs a day.
View the post on LinkedIn
Garry TanCaptain is a step function increase in capability compared to out-of-box RAG pipelines.
View the post on LinkedInCaptain queries storage buckets and cloud drives with state-of-the-art accuracy, and queries are free. Public benchmarks are a starting point: run with Captain’s MCP to measure retrieval on your own files.
When evaluated against other leading file-search APIs on production queries, alternatives fail to answer about one in three questions.
Every answer comes with the exact snippet the agent saw: a PDF region, an audio or video clip, a spreadsheet’s rows and columns. Verifying a response is easy.
Start at $0 a month, and queries are free. Use the API, or the MCP server from the coding agent you already work in.

Our famous file search API.
Point it at your files and build on it. Parsing, embeddings, vector DB and evaluation are the infrastructure underneath.

Evaluate and tune your file search.
Call captain_eval from the hosted MCP server and it scores retrieval on evals written from your own files, then tunes toward the accuracy or latency target you set.

Bucket change notifications drive the index in near real time, and scheduled reconciliation jobs sweep for anything a notification dropped.
Your buckets remain the source of truth.
Set the rules for relevant results. Filters and ranking controls help match each query with the right information.
Captain’s AI agents find search problems and test improvements against your own files and queries. Your team stays in control.
We do the heavy lifting for your compliance team. Your files are encrypted in transit and at rest.
Opt-in PII masking redacts sensitive information when your files are indexed, so it can be hidden from your AI agents.
Search content inside scanned documents, tables and recordings without building a parsing pipeline.
XML fragments in a zip, with tables that only exist as formatting. Captain's parser was tuned on the ugliest DOCX files our customers could find.
Clean PDFs parse anywhere. Captain also reads the rotated scan and the handwritten margin note.
Transcribed, described and indexed at ingest. The chart on slide 12 is a query, not a scrubbing session.
TLS in transit and encryption at rest protect your data. Opt-in PII masking helps keep sensitive file content out of search results.