A plugin that prunes OpenClaw's 40K-token workspace injection down to only the files relevant to each message, cutting baseline API costs by 60-80%
OpenClaw injects the entire workspace context (SOUL.md, agent configs, skill descriptions, tool declarations) into every single message, consuming 40,000+ tokens before the user even types a word. This fixed overhead makes smaller models unusable (they hit context limits immediately) and multiplies API costs 3-5x for conversational sessions. This plugin intercepts the context assembly step and applies relevance filtering, only including workspace files and tool declarations that match the current message's intent, dropping the baseline from 40K to under 10K tokens.
Demand Breakdown
Social Proof 3 sources
Gap Assessment
4 tools exist (lossless-claw, OpenClaw-Token-Optimizer, ClawRouter, maxHistoryMessages (built-in)) but gaps remain: Only compresses history, does not touch baseline workspace injection or tool declaration bloat. The 40K baseline persists even with lossless-claw active.; Works at the routing and scheduling level, not at the context assembly level. Cannot prune workspace files or tool declarations from the prompt..
Features4 agent-ready prompts
Competitive LandscapeFREE
| Product | Does | Missing |
|---|---|---|
| lossless-claw | Compresses conversation history without losing critical information via a Context Engine plugin | Only compresses history, does not touch baseline workspace injection or tool declaration bloat. The 40K baseline persists even with lossless-claw active. |
| OpenClaw-Token-Optimizer | General token optimization skill combining model routing, heartbeat tuning, and usage tracking | Works at the routing and scheduling level, not at the context assembly level. Cannot prune workspace files or tool declarations from the prompt. |
| ClawRouter | Smart LLM router that reduces costs by routing simple tasks to cheaper models | Routing alone doesn't fix baseline bloat. A 40K-token baseline still goes to the cheap model, just at a lower per-token rate. The waste is the same tokens, just cheaper. |
| maxHistoryMessages (built-in) | Limits conversation history to N messages to reduce context growth over time | Only limits history messages, not the system prompt or workspace injection. Baseline stays at 40K+ regardless of history settings. |
Sign in to unlock full access.