docs: reference compressor-reflex-mcp PyPI package on model card
Browse filesAdds a short MCP server subsection linking the PyPI package that deploys this model in IDEs.
README.md
CHANGED
|
@@ -44,6 +44,10 @@ model-index:
|
|
| 44 |
- **Decision rule:** keep line iff P(keep) ≥ τ\*, with τ\*=0.50 calibrated on a held-out must-keep anchor set.
|
| 45 |
- **Fail-open policy:** outputs of ≤5 physical lines or ≤64 tokens bypass the compressor and pass through verbatim; if scoring ever yields zero kept lines, the first and last lines are kept. The model never silently drops everything.
|
| 46 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
## Intended Uses & Limitations
|
| 48 |
|
| 49 |
**Intended use:** compressing tool outputs (build logs, test results, grep/file reads, command output) before they enter an LLM's context window, in agentic coding workflows. Designed as a local MCP-side layer.
|
|
|
|
| 44 |
- **Decision rule:** keep line iff P(keep) ≥ τ\*, with τ\*=0.50 calibrated on a held-out must-keep anchor set.
|
| 45 |
- **Fail-open policy:** outputs of ≤5 physical lines or ≤64 tokens bypass the compressor and pass through verbatim; if scoring ever yields zero kept lines, the first and last lines are kept. The model never silently drops everything.
|
| 46 |
|
| 47 |
+
### MCP server
|
| 48 |
+
|
| 49 |
+
The companion MCP server packages this model for IDE deployment (Cursor, Antigravity, Claude Desktop): install with `pip install compressor-reflex-mcp` ([PyPI](https://pypi.org/project/compressor-reflex-mcp/)). It downloads these weights from this repo on first run and exposes `compress_tool_output` / `compress_file` tools plus a transparent proxy mode.
|
| 50 |
+
|
| 51 |
## Intended Uses & Limitations
|
| 52 |
|
| 53 |
**Intended use:** compressing tool outputs (build logs, test results, grep/file reads, command output) before they enter an LLM's context window, in agentic coding workflows. Designed as a local MCP-side layer.
|