SeeLLM documentation

Start with the work you need to complete.

SeeLLM documents evidence separately from inference so people and automation can make decisions from the same verifiable record.

For people

Evaluate

Choose important pages and establish what a public check can actually show.

Install

Connect supported server-side observation and confirm adapter activity.

Interpret

Separate observed, sampled, and inferred evidence before acting.

Weekly workflow

Review request changes, choose a bounded follow-up, and recheck the same context.

For agents & automation

MCP

Configure a Streamable HTTP client and verify the available tools.

REST

Make a first authenticated request against a documented endpoint.

Jobs

Read evidence, distinguish implementation from work, and record an action.

Agent Hooks

Verify signed job deliveries before an agent accepts work.