imux · 龍imux
AI command center for macOS

Errors & support

When Chat, the API, or imux official models fail, you get a readable message, a stable error code, and a support ref (requestId). This page shows how to diagnose quickly and how to report issues.

Start with the Ref line at the end of the error (e.g. mbr_xxx · upstream_error · stage=upstream · HTTP 502). Paste that line when contacting support so we can correlate server logs.

Where errors appear

  • imux 原生 App: Agent Chat bubble shows human copy + Ref; Console has [AgentChatProvider] logs.
  • Web Chat / Playground/playground · chat.clavue.com): red error banner + Ref; browser console logs [clavue-chat] membership error.
  • APIapi.clavue.com/v1/* / www.clavue.com/api/membership/v1/*): JSON envelope + headers x-imux-request-id / x-imux-error-code / x-imux-error-stage.

How to read the support Ref

Typical format:

Official model upstream failed. Retry in a moment, or switch model (e.g. clavue-2.1-fast).
Slow down and retry later.

Ref: mbr_mrz… · upstream_error · stage=upstream · HTTP 502 · up=500
PartMeaning
mbr_…requestId — correlates one request end-to-end
upstream_errorstable error code — drives recovery
stage=upstreampipeline stage: auth / quota / upstream / validation…
HTTP 502HTTP status returned to you (may be mapped, e.g. upstream 401→503)
up=500raw upstream status when available

30-second triage

  1. Confirm you are signed in (Account shows email / plan).
  2. Read the code: quota / rate limit → Account & Pricing; upstream_* → retry or switch model.
  3. Try another official model (auto ↔ clavue-2.1-fast).
  4. Still failing: copy the full error (including Ref) and report.

Common error codes

codeHTTPMeaningWhat you can do
sign_in_required401Not signed in, or session/token expiredOpen Account and sign in again; CLI: re-run device login
rate_limited429Too many requests (IP or plan RPM)Wait and retry; upgrade plan for higher RPM
quota_exceeded402Daily / weekly / monthly or boost quota emptyUpgrade, buy boost packs, check-in, or wait for reset
official_models_only400Non-official model id in the requestUse auto / clavue / clavue-2.1 / -fast / -pro / -rev
upstream_error502Official model upstream failed (e.g. openai_error)Retry; switch model (e.g. clavue-2.1-fast); if persistent, report with Ref
upstream_rate_limited429Upstream provider is rate-limitingSlow down and retry later
upstream_not_configured503Service misconfigured (ops side)Retry later; report with requestId if it lasts
upstream_model_unavailable502Mapped upstream model not availableSwitch product model; report if all models fail
context_length_exceeded400Prompt / history too long for contextStart a new chat or shorten the request
internal_error500Unexpected server errorRetry; report with requestId if it repeats

Full developer envelope: Developers and repo docs/membership-api-errors.md

Signed in but still failing? Not always auth

Older opaque tokens like openai_error looked like “not signed in”. Now:

  • True auth failure → code=sign_in_required, stage=auth
  • Official upstream down → code=upstream_*, stage=upstream (turn quota refunded)
  • Out of quota → code=quota_exceeded, stage=quota

Quota & boost packs

  • Usage: /account · /usage
  • Upgrade: /pricing
  • Boost: check-in / invite / rush on Account; auto-spend after free quota.

How to report a problem

Include the following so we can jump to the requestId in one step:

  1. Full error text (including the Ref line)
  2. Where: imux app / Web Chat / API / CLI
  3. Approx. time (with timezone) and model id
  4. Repro steps; whether switching model or a new chat helps
  5. API users: x-imux-request-id header + response JSON
Do not paste API keys, passwords, full session cookies, or private chat content. requestId + codes are enough.

Channels: GitHub Issues · in-app feedback when available · community board.

API self-check

# Expect 401 + requestId + code=sign_in_required when logged out
curl -si https://api.clavue.com/v1/chat/completions \
  -H "content-type: application/json" \
  -H "x-imux-request-id: mbr_cli_demo_001" \
  -d '{"model":"auto","messages":[{"role":"user","content":"hi"}]}' \
  | head -40

# Signed-in call (replace TOKEN)
curl -si https://api.clavue.com/v1/chat/completions \
  -H "Authorization: Bearer $TOKEN" \
  -H "content-type: application/json" \
  -d '{"model":"auto","messages":[{"role":"user","content":"hi"}]}'