graff — a fast agentic coding harness in Zig: multi-provider, MCP, workflows, DGM evolution loop, TS/Python SDKs
Python
v0.302.48
- oMLX is a built-in local provider.
--model omlxtalks to the oMLX app's OpenAI-compatible server on127.0.0.1:8000from any directory, with no.graff/.config.router. Store its key withgraff key set omlx <key>or setOMLX_API_KEY.graff modelslists the models oMLX serves, a bare--model omlxstarts on the first one,/effortand ACP thought levels are sent asreasoning_effort, and calls are billed as local (#1695). - Local model windows come from the server. Live model catalogs read
max_model_len, which oMLX and vLLM-style servers report, so a local model gets its real context window instead of the 200k fallback. - A stopped local server is named. When a loopback server refuses the connection, the failed turn ends with the URL it tried.
- Compaction counts opaque state by the provider's own measure. Compaction state was estimated at bytes / 16, so a large blob could read as several times its real size and keep the context over the compaction point, compacting every turn (#1686). After a full request that carries the state, its share of the provider's reported input replaces the estimate for that state.
- Retains all changes from v0.302.47.