open source
The official clients for the Codag action API, plus the versioned OpenAPI spec they are built from.
Reduce action results, read contentless usage and savings, and manage workspace policy from your own applications, harnesses, and CI. One idiomatic client per language, same contract everywhere.
$ pip install codag $ npm install @codag/sdk $ go get github.com/codag-megalith/codag-sdk/go one call: reduce_action(envelope) decision: reduced | passthrough
Deterministic log templating: turns raw log lines into template groups with counts, no model involved.
A Drain-based grouper with adaptations for compact, punctuation-heavy machine logs. Bounded raw samples, slot summaries, fully reproducible output. It came out of Codag's log work and stays open and maintained.
$ codag-drain < app.log 1,240,000 lines 14 templates [812,394] GET /v1/users/<*> 200 [404,001] cache <*> user:<*> [ 23,118] pool acquire <*>ms [ 412] ERROR psycopg2.OperationalError
See your AI workflows as live diagrams, right inside VS Code. Works with any language, model, or framework.
Edit code and the graph refreshes live. Click a node to jump to source. Tree-sitter parsing under the hood, no config required.
Signed release artifacts, checksums, provenance, and attestations for the Codag binary are published in codag-releases. The install script at codag.ai/install.sh resolves to them directly.