FAQ
Account
Verification email didn't arrive?
- Check spam / Promotions
- QQ / 163 mail occasionally blocks; prefer Gmail / Outlook / Apple iCloud
- Try Forgot password on the login page to confirm deliverability
Forgot password?
Use Forgot password on the login page, or sign in via Passkey.
No third-party sign-in?
Not right now — no GitHub / Discord OAuth. Email sign-up + Passkey + 2FA only.
Lost Passkey?
Bridge binds one Passkey per account, so losing the device means that Passkey is gone. What to do:
- Sign in with username / email + password, then complete 2FA (TOTP)
- A backup code works in place of TOTP for a one-shot login
- If none of the above works, email
[email protected]from the registered address and ask the admin to reset your account manually
Strongly recommended
Enable 2FA and save the backup codes at the same time you bind a Passkey — otherwise losing the device plus forgetting the password leaves only the manual-reset path.
API Keys
Key format?
sk-…. Shown once at creation — copy immediately.
Lost the key / forgot to copy?
Can't be recovered. Delete it in API Keys, create a new one.
Suspect a leak?
Disable or delete the key on API Keys immediately. Check the last 24h of Usage Logs for abuse, email [email protected] if needed.
How many keys per account?
No hard cap in practice. Each key supports quota / IP / model whitelist / expiry independently.
Billing
Where do I check prices?
Model Catalog. All numbers are final user-facing prices.
How long does $1 last?
Depends on model and prompt length. Top up $5 first and observe your actual burn rate.
A call charged more / less than expected?
Usage Logs has the per-call token breakdown.
How does cache billing work?
Each model card has three columns: Input / Output / Cache. Cache hits bill at the cache price, usually much lower than input.
How do I claim the daily check-in?
Profile → Daily check-in button.
Are there redemption codes / referrals?
Yes. Wallet has a "Got a redemption code?" input and a Referral program section with your personal link.
Models
Model X is missing?
Use Filter or the search box on Model Catalog. If still nothing:
- Possibly deprecated
- Possibly only in certain groups — check the model card's Group tag
- Possibly not yet configured — email to request
Calling model X returns 403?
Your key's group doesn't include it. Make a new key in a group that does.
Can I call Claude / Gemini via /v1/chat/completions?
Yes, Bridge auto-translates. But for full fidelity (Claude tool-use schema, Gemini grounding), use the native protocol.
Integration
My client doesn't allow a custom base URL?
- Some clients hide it under "Advanced". Look there first.
- If genuinely unsupported: switch to a supported client — Cherry Studio, DeepChat, Lobe Chat, etc. API Keys has one-click import buttons.
Base URL set but calls fail?
Walk through:
- URL typos — no trailing
/; OpenAI uses/v1, Anthropic doesn't add/v1, Gemini uses/gemini - Whitespace / newline from copy-paste
- Usage Logs has this call? If not, request never hit Bridge.
- If yes, check the
upstream_errorfield
Stability
Most stable group?
default / anthropic / openai / google go direct to upstream — most stable. Discount groups (cc / aws-q / codex / windsurf) re-use subscriptions — better value but occasional upstream waves. For critical flows, keep two keys — one official, one discounted.
Upstream 429?
Bridge buffers some of this, but extreme bursts will pass through. Implement exponential backoff client-side.
Other
User Agreement / Privacy Policy?
Support?
Email [email protected].