Every tool call
becomes model input.
LeanCTX normalizes, compresses and audits the tool output your automations feed to models: 95+ shell patterns, provider connectors for GitHub, GitLab, Jira and Postgres, composable context via ctx_compose, and a complete audit trail. Pipelines get cheaper, deterministic and reviewable.
What it costs you today.
CI logs torch your budget
A failing pipeline dumps megabytes of logs into your triage agent. The signal is three lines; you pay for all of them.
Each integration is bespoke
Jira here, GitHub there, a database somewhere. Every workflow glues its own fetch-and-paste code with no shared layer.
Nobody can audit what agents saw
When an automation acts on stale or wrong context, you need to know exactly what it read. Most stacks can’t answer that.
The capabilities that do the work.
Everything below ships in the open-source binary today. No roadmap items, no waitlists.
From zero to first gain.
One guide. Two journeys. Full reference.
Questions teams ask before adopting.
How much smaller does shell output get?
88–99% on common tools: a cargo build shrinks from hundreds of lines to errors plus result, npm install to the dependency delta. 95+ patterns ship today, and unknown commands fall back to safe generic compression.
Can LeanCTX pull context from our internal systems?
Yes. Providers for GitHub, GitLab, Jira and Postgres ship today, plus a custom REST provider. All provider data flows through the same consolidation pipeline into search, graph and knowledge.
Is automation output auditable?
Every read, command and saving lands in a local, Ed25519-signed ledger. You can export exactly what an automation consumed and verify the chain.
Take back control of your context.
Free for local use, forever. CI enforces it. One binary, ten minutes to the first measured gain.