pi-mono/packages/ai/scripts
Armin Ronacher 962b130e0c feat(coding-agent): add in-context compaction
With compaction.inContext enabled, compaction can generate the summary inside the cached conversation: it repeats the request the next turn would send, appends the instructions as a system message plus a user message, and sets toolChoice none. The recent tail is kept as before. A gate requires a verified model, a warm cache from the same model and thinking level, room in the context window, and a lower estimated cost than the standalone summary. Failed or rejected attempts fall back to the standalone summary, never execute tools, and are recorded in details.inContext with their usage.

pi-ai adds the supportsInContextCompaction compat setting and test/in-context-compaction-probe.ts, which measures whether a model keeps its prompt cache for this request shape. The generated catalog flags the models that passed.
2026-10-07 00:14:46 +02:00
..
check-model-data.ts fix(ai): validate generated model data before builds 2026-07-20 22:29:06 +02:00
generate-models.ts feat(coding-agent): add in-context compaction 2026-10-07 00:14:46 +02:00
generate-test-image.ts chore: pin dependencies and use native TypeScript 2026-05-20 12:46:17 +02:00
hydrate-model-catalog.ts feat(coding-agent): add Nix flake (#9137) 2026-10-02 18:30:51 +02:00
model-data.ts feat(coding-agent): add Nix flake (#9137) 2026-10-02 18:30:51 +02:00
models-dev-reasoning-options.ts generate-models: use reasoning options from models.dev (#6928) 2026-07-22 08:33:34 +02:00
openrouter-catalog.ts feat(ai): serve Jev classifier through OpenRouter and Cloudflare Workers AI 2026-09-24 12:34:58 +02:00
openrouter-reasoning-options.ts fix(ai): derive OpenRouter reasoning controls (#8614) 2026-08-25 12:11:21 +02:00