openclaw/docs/reference/token-use.md

summary, read_when, title

summary	read_when	title
How OpenClaw builds prompt context and reports token usage + costs	Explaining token usage, costs, or context windows Debugging context growth or compaction behavior	Token use and costs

OpenClaw tracks tokens, not characters. Tokens are model-specific, but most OpenAI-style models average ~4 characters per token for English text.

How the system prompt is built

OpenClaw assembles its own system prompt on every run. It includes:

• Tool list + short descriptions
• Skills list (only metadata; instructions are loaded on demand with read). The compact skills block is bounded by skills.limits.maxSkillsPromptChars, with optional per-agent override at agents.list[].skillsLimits.maxSkillsPromptChars.
• Self-update instructions
• Workspace + bootstrap files (AGENTS.md, SOUL.md, TOOLS.md, IDENTITY.md, USER.md, HEARTBEAT.md, BOOTSTRAP.md when new, plus MEMORY.md when present). Lowercase root memory.md is not injected; it is legacy repair input for openclaw doctor --fix when paired with MEMORY.md. Large files are truncated by agents.defaults.bootstrapMaxChars (default: 12000), and total bootstrap injection is capped by agents.defaults.bootstrapTotalMaxChars (default: 60000). memory/*.md daily files are not part of the normal bootstrap prompt; they remain on-demand via memory tools on ordinary turns, but reset/startup model runs can prepend a one-shot startup-context block with recent daily memory for that first turn. Bare chat /new and /reset commands are acknowledged without invoking the model. The startup prelude is controlled by agents.defaults.startupContext.
• Time (UTC + user timezone)
• Reply tags + heartbeat behavior
• Runtime metadata (host/OS/model/thinking)

See the full breakdown in System Prompt.

What counts in the context window

Everything the model receives counts toward the context limit:

• System prompt (all sections listed above)
• Conversation history (user + assistant messages)
• Tool calls and tool results
• Attachments/transcripts (images, audio, files)
• Compaction summaries and pruning artifacts
• Provider wrappers or safety headers (not visible, but still counted)

Some runtime-heavy surfaces have their own explicit caps:

• agents.defaults.contextLimits.memoryGetMaxChars
• agents.defaults.contextLimits.memoryGetDefaultLines
• agents.defaults.contextLimits.toolResultMaxChars
• agents.defaults.contextLimits.postCompactionMaxChars

Per-agent overrides live under agents.list[].contextLimits. These knobs are for bounded runtime excerpts and injected runtime-owned blocks. They are separate from bootstrap limits, startup-context limits, and skills prompt limits.

For images, OpenClaw downscales transcript/tool image payloads before provider calls. Use agents.defaults.imageMaxDimensionPx (default: 1200) to tune this:

• Lower values usually reduce vision-token usage and payload size.
• Higher values preserve more visual detail for OCR/UI-heavy screenshots.

For a practical breakdown (per injected file, tools, skills, and system prompt size), use /context list or /context detail. See Context.

How to see current token usage

Use these in chat:

• /status → emoji-rich status card with the session model, context usage, last response input/output tokens, and estimated cost (API key only).
• /usage off|tokens|full → appends a per-response usage footer to every reply.
  • Persists per session (stored as responseUsage).
  • OAuth auth hides cost (tokens only).
• /usage cost → shows a local cost summary from OpenClaw session transcripts.

Other surfaces:

• TUI/Web TUI: /status + /usage are supported.
• CLI: openclaw status --usage and openclaw channels list show normalized provider quota windows (X% left, not per-response costs). Current usage-window providers: Anthropic, GitHub Copilot, Gemini CLI, OpenAI Codex, MiniMax, Xiaomi, and z.ai.

Usage surfaces normalize common provider-native field aliases before display. For OpenAI-family Responses traffic, that includes both input_tokens / output_tokens and prompt_tokens / completion_tokens, so transport-specific field names do not change /status, /usage, or session summaries. Gemini CLI JSON usage is normalized too: reply text comes from response, and stats.cached maps to cacheRead with stats.input_tokens - stats.cached used when the CLI omits an explicit stats.input field. For native OpenAI-family Responses traffic, WebSocket/SSE usage aliases are normalized the same way, and totals fall back to normalized input + output when total_tokens is missing or 0. When the current session snapshot is sparse, /status and session_status can also recover token/cache counters and the active runtime model label from the most recent transcript usage log. Existing nonzero live values still take precedence over transcript fallback values, and larger prompt-oriented transcript totals can win when stored totals are missing or smaller. Usage auth for provider quota windows comes from provider-specific hooks when available; otherwise OpenClaw falls back to matching OAuth/API-key credentials from auth profiles, env, or config. Assistant transcript entries persist the same normalized usage shape, including usage.cost when the active model has pricing configured and the provider returns usage metadata. This gives /usage cost and transcript-backed session status a stable source even after the live runtime state is gone.

OpenClaw keeps provider usage accounting separate from the current context snapshot. Provider usage.total can include cached input, output, and multiple tool-loop model calls, so it is useful for cost and telemetry but can overstate the live context window. Context displays and diagnostics use the latest prompt snapshot (promptTokens, or the last model call when no prompt snapshot is available) for context.used.

Cost estimation (when shown)

Costs are estimated from your model pricing config:

models.providers.<provider>.models[].cost

These are USD per 1M tokens for input, output, cacheRead, and cacheWrite. If pricing is missing, OpenClaw shows tokens only. OAuth tokens never show dollar cost.

After sidecars and channels reach the Gateway ready path, OpenClaw starts an optional background pricing bootstrap for configured model refs that do not already have local pricing. That bootstrap fetches remote OpenRouter and LiteLLM pricing catalogs. Set models.pricing.enabled: false to skip those catalog fetches on offline or restricted networks; explicit models.providers.*.models[].cost entries continue to drive local cost estimates.

Cache TTL and pruning impact

Provider prompt caching only applies within the cache TTL window. OpenClaw can optionally run cache-ttl pruning: it prunes the session once the cache TTL has expired, then resets the cache window so subsequent requests can re-use the freshly cached context instead of re-caching the full history. This keeps cache write costs lower when a session goes idle past the TTL.

Configure it in Gateway configuration and see the behavior details in Session pruning.

Heartbeat can keep the cache warm across idle gaps. If your model cache TTL is 1h, setting the heartbeat interval just under that (e.g., 55m) can avoid re-caching the full prompt, reducing cache write costs.

In multi-agent setups, you can keep one shared model config and tune cache behavior per agent with agents.list[].params.cacheRetention.

For a full knob-by-knob guide, see Prompt Caching.

For Anthropic API pricing, cache reads are significantly cheaper than input tokens, while cache writes are billed at a higher multiplier. See Anthropic's prompt caching pricing for the latest rates and TTL multipliers: https://docs.anthropic.com/docs/build-with-claude/prompt-caching

Example: keep 1h cache warm with heartbeat

agents:
  defaults:
    model:
      primary: "anthropic/claude-opus-4-6"
    models:
      "anthropic/claude-opus-4-6":
        params:
          cacheRetention: "long"
    heartbeat:
      every: "55m"

Example: mixed traffic with per-agent cache strategy

agents:
  defaults:
    model:
      primary: "anthropic/claude-opus-4-6"
    models:
      "anthropic/claude-opus-4-6":
        params:
          cacheRetention: "long" # default baseline for most agents
  list:
    - id: "research"
      default: true
      heartbeat:
        every: "55m" # keep long cache warm for deep sessions
    - id: "alerts"
      params:
        cacheRetention: "none" # avoid cache writes for bursty notifications

agents.list[].params merges on top of the selected model's params, so you can override only cacheRetention and inherit other model defaults unchanged.

Example: enable Anthropic 1M context beta header

Anthropic's 1M context window is currently beta-gated. OpenClaw can inject the required anthropic-beta value when you enable context1m on supported Opus or Sonnet models.

agents:
  defaults:
    models:
      "anthropic/claude-opus-4-6":
        params:
          context1m: true

This maps to Anthropic's context-1m-2025-08-07 beta header.

This only applies when context1m: true is set on that model entry.

Requirement: the credential must be eligible for long-context usage. If not, Anthropic responds with a provider-side rate limit error for that request.

If you authenticate Anthropic with OAuth/subscription tokens (sk-ant-oat-*), OpenClaw skips the context-1m-* beta header because Anthropic currently rejects that combination with HTTP 401.

Tips for reducing token pressure

• Use /compact to summarize long sessions.
• Trim large tool outputs in your workflows.
• Lower agents.defaults.imageMaxDimensionPx for screenshot-heavy sessions.
• Keep skill descriptions short (skill list is injected into the prompt).
• Prefer smaller models for verbose, exploratory work.

See Skills for the exact skill list overhead formula.

Related

• API usage and costs
• Prompt caching
• Usage tracking