solution install high linux macos windows

Onboarding skips Model/Auth setup (agent unresponsive after install)

Fix a broken onboarding flow where OpenClaw defaults to a model without credentials, causing the agent to hang, by configuring provider auth manually and rerunning current configuration.

By CoClaw Team • • Updated June 8, 2026

Symptoms

  • openclaw onboard finishes, but you were never asked to pick a model provider or enter an API key.
  • OpenClaw defaults to anthropic/claude-opus-4-6, but the agent never responds (TUI/WebChat shows “loading” indefinitely).
  • Re-running openclaw onboard does not bring back the missing Model/Auth step.

Cause

The onboarding wizard can be broken in some releases and skip the Model/Auth step entirely. When that happens, OpenClaw may default to a provider/model that requires credentials, but no credentials were configured, so the agent cannot complete requests.

Fix

1) Configure a provider API key after onboarding (fastest)

If you’re using Anthropic:

openclaw config set env.ANTHROPIC_API_KEY "sk-ant-..."
openclaw gateway restart

If you’re using OpenAI instead, set env.OPENAI_API_KEY (or configure it via openclaw configure) and restart the gateway.

2) Run the interactive configurator

openclaw configure
openclaw gateway restart

3) Recover after a release-specific onboarding regression

If possible, move back to the current stable path, rerun configuration, and restart the gateway:

openclaw configure
openclaw gateway restart

If you are responding to a proven release-specific regression and have a personally validated last-working version, follow the generic rollback flow in /guides/updating-and-migration. Do not use a historical pin by default.

Verify

  • Send a message in TUI/WebChat; the agent responds normally.
  • Optional: openclaw models status --probe succeeds for your configured provider.
  • GitHub issue: #16134

Verification & references

  • Reviewed by:CoClaw Code Team
  • Last reviewed:June 8, 2026
  • Verified on: Linux · macOS · Windows

References

Official docs

  1. OpenClaw docs: /config
  2. OpenClaw docs: /start/wizard

Issue threads

  1. OpenClaw GitHub issue #16134
Want to explore more? Browse all solutions or ask in the Community Forum .
Report a problem

Related Resources

Model/auth failures: rate limit, billing, or 'all models failed'
Fix
Debug OpenClaw model failures by checking provider auth status, probing profiles, switching models/fallbacks, and verifying provider/model refs.
Ollama configured, but OpenClaw still uses Anthropic (or model discovery keeps failing)
Fix
Fix local Ollama setups where gateway logs show Anthropic fallback or repeated Ollama model-discovery failures by pinning provider config, verifying connectivity from the gateway runtime, and separating model selection problems from OpenAI-compatible payload problems.
/model says 'model not allowed'
Fix
Fix 'model not allowed' by updating or clearing the agents.defaults.models allowlist, then selecting an allowed model.
OpenClaw config becomes invalid after you remove a provider but keep its auth profile
Fix
Fix config validation failures caused by orphaned `auth.profiles` entries that still reference a provider removed from `models.providers`.
OpenClaw Safe Mode Recovery After OAuth Renewal: auth-profiles.json vs provisioned copy
Guide
Understand why OpenClaw recovery can keep looping after a successful OAuth renewal, and how to resync the live auth file, the provisioned copy, and safe-mode markers.
OpenClaw Email (Gmail/IMAP) Setup: OAuth Reliability, Re-Auth Loops, and Safer Alternatives
Guide
Set up OpenClaw email with a path you can actually keep stable: choose the right auth model, reduce Gmail re-auth loops, contain blast radius, and verify reliability before the inbox becomes business-critical.