Skip to main content
Caesar is a search API built for agents. Every result carries stable handles — doc_id, passage_id, search_id — that you can read in full, cite, and feed back. No setup required: the anonymous tier is live. A surreal animated collage of a hand holding a glowing glass orb while graph-paper cards drift in a seamless loop.

​
Connect an agent first

For agents that support skills, install Caesar search directly: 
npx skills add caesar-data/skills --skill caesar-search

Use Caesar to search, read the most relevant result, cite the source URL, and preserve doc_id and search_id.

For agents: use /llms.txt to discover docs, append .md to any page for markdown, and start with /agents/install for client-specific setup.

​
Or call the API directly

This works as written with only curl: 
curl -s https://search-api-staging-779189860552.europe-west1.run.app/v1/search \
  -H "Content-Type: application/json" \
  -d '{"query": "postgres 17 logical replication failover", "max_results": 3}'
Each result includes a doc_id you can pass to /v1/document to read the page in full, and the response includes a search_id you can pass to /v1/feedback to tell Caesar what helped. Run the full loop in 60 seconds.

​
The product loop

Search

Get ranked documents with search_id and doc_id.

Read

Read the selected doc_id as markdown.

Cite

Preserve source URL, passage, capture, and time.

Feedback

Send what helped back with search_id and doc_id.

​
Three verbs, every surface

VerbEndpointWhat it does
searchPOST /v1/searchRanked retrieval over canonical documents and passages
readPOST /v1/documentInspect one document and retrieve its content (full-page markdown with content.selection: "full_document")
feedbackPOST /v1/feedbackRecord what helped, so ranking improves
The same three verbs appear on every surface: search/read/feedback in the CLI and SDKs, caesar_search/caesar_read over MCP. Learn one surface and you know them all.

​
Provenance is the point

Search results from most APIs are dead ends: a URL and a snippet. Caesar results are live handles:
  • doc_id — deterministic identifier for a canonical document, stable across searches and recrawls. Read it in full any time with /v1/document.
  • passage_id — handle to a specific passage in the latest capture, with section context.
  • search_id — handle to the search itself, used to attribute feedback to the exact ranking that produced a result.
  • capture_id and capture_time — exactly which capture of the document your content came from (on search results with response.verbosity: "full"; always available from /v1/document).
That means an agent can search, quote a passage, cite where and when it was captured, and report back which results were worth the tokens. See provenance for the full object model.

​
Pick a surface

API reference

Three endpoints, an OpenAPI spec, and an interactive playground. Start here for raw HTTP.

How search works

Canonical documents, passages, freshness, ranking, and provenance handles.

Pricing

Free to use. Anonymous access works without a key; partner API keys add higher throughput.

Python SDK

pip install caesar-search — sync and async clients, typed models, automatic retries.

TypeScript SDK

npm install caesar-search — fully typed, ESM and CJS, works in Node, Bun, Deno, and edge runtimes.

AI SDK tools

caesar_search and caesar_read as drop-in tools from the caesar-search/ai subpath.

CLI

caesar-search — search, read, and feedback from the shell, with a --json contract built for scripts and agents.

MCP server

Remote MCP endpoint at /mcp — two tools, no local install, one command to connect.

Integrations

Recipes for framework tools, generic tool calling, and Agent Development Kit projects.

Agent Skills

Install the search skill with npx skills add, or use the API installer for the full bundle.

​
Next steps

  • Quickstart — search, read, and feedback with nothing but curl.
  • How search works - how Caesar turns public pages into citable documents and passages.
  • Pricing - free access, anonymous tier, and partner keyed throughput.
  • Authentication — what an API key changes, and why you may not need one yet.
  • Agents guide — conventions for building agents on Caesar.