#Too much token usage

1 messages · Page 1 of 1 (latest)

fossil tuskBOT
#

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> âś… Mark Solution

sullen orchidBOT
#

Your message got cut off at “The frist…” — can you paste the rest of that sentence (or a screenshot of what happens on the first step)?

In the meantime, since you’re in #too-much-token-usage, the most common “first message burns a ton of tokens” causes are:

  1. Huge bootstrap files getting injected on every run (AGENTS.md / SOUL.md / USER.md / MEMORY.md, etc.).
  2. You accidentally enabled long-context (1M) mode (context1m: true) and Claude rejects it (or it’s just very expensive).

Quick checks (copy/paste results, redact secrets)

In chat (best):

  • /status
  • /context detail (this shows what’s actually consuming tokens)
  • /usage tokens (so replies show token footers)

On the gateway host:

openclaw --version