Skip to content

Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.

feat(cli): add IntelligentContext CLI arguments

chopratejasd57673d745e9
diff --git a/README.md b/README.md
index 77f865d40..15ce7cb41 100644
--- a/README.md
+++ b/README.md
@@ -146,7 +146,7 @@ flowchart TB
subgraph Pipeline["Transform Pipeline"]
CA["Cache Aligner<br/><i>Stabilizes dynamic tokens</i>"]
SC["Smart Crusher<br/><i>Removes redundant tool output</i>"]
- CM["Context Manager<br/><i>Fits token budget</i>"]
+ CM["Intelligent Context<br/><i>Score-based token fitting</i>"]
CA --> SC --> CM
end
@@ -278,7 +278,7 @@ See the full [Agno Integration Guide](docs/agno.md) for hooks, multi-provider su
| **Universal Compression** | ML-based content detection + structure-preserving compression | [Compression](docs/compression.md) |
| **SmartCrusher** | Compresses JSON tool outputs statistically | [Transforms](docs/transforms.md) |
| **CacheAligner** | Stabilizes prefixes for provider caching | [Transforms](docs/transforms.md) |
-| **RollingWindow** | Manages context limits without breaking tools | [Transforms](docs/transforms.md) |
+| **IntelligentContext** | Score-based context dropping with TOIN-learned importance | [Transforms](docs/transforms.md) |
| **CCR** | Reversible compression with automatic retrieval | [CCR Guide](docs/ccr.md) |
| **LangChain** | Memory, retrievers, agents, streaming | [LangChain](docs/langchain.md) |
| **Agno** | Agent framework integration with hooks | [Agno](docs/agno.md) |
diff --git a/headroom/cli.py b/headroom/cli.py
index 44d5e0e80..e2f1b5194 100644
--- a/headroom/cli.py
+++ b/headroom/cli.py
@@ -67,6 +67,10 @@ def cmd_proxy(args: argparse.Namespace) -> int:
llmlingua_target_rate=args.llmlingua_rate,
# Code-aware: ON by default (use --no-code-aware to disable)
code_aware_enabled=not args.no_code_aware,
+ # Intelligent Context: ON by default (use --no-intelligent-context to disable)
+ intelligent_context=not args.no_intelligent_context,
+ intelligent_context_scoring=not args.no_intelligent_scoring,
+ intelligent_context_compress_first=not args.no_compress_first,
# Memory System
memory_enabled=args.memory,
memory_backend=args.memory_backend,
@@ -491,6 +495,22 @@ Documentation: https://github.com/headroom-sdk/headroom
action="store_true",
help="Disable AST-based code compression",
)
+ # Intelligent Context Management (ON by default)
+ proxy_parser.add_argument(
+ "--no-intelligent-context",
+ action="store_true",
+ help="Disable IntelligentContextManager (fall back to RollingWindow)",
+ )
+ proxy_parser.add_argument(
+ "--no-intelligent-scoring",
+ action="store_true",
+ help="Disable multi-factor importance scoring (use position-based)",
+ )
+ proxy_parser.add_argument(
+ "--no-compress-first",
+ action="store_true",
+ help="Disable trying deeper compression before dropping messages",
+ )
# Memory System
proxy_parser.add_argument(
"--memory",