Skip to main content

​
Error Reference

This page documents the error responses returned by the thig.ai API, including HTTP status codes, application-specific error codes, and troubleshooting tips.

​
Error Response Format

All API errors return a JSON body with a human-readable message and an optional error code:
The error field is always present. The code field is included when a specific application error code applies. Some endpoints may include additional fields such as details for validation errors.

​
HTTP Status Codes

​
Error Codes

​
SSE Error Events

Streaming endpoints (such as Chat and PRD generation) use Server-Sent Events (SSE) to deliver responses incrementally. When an error occurs during streaming, it is sent as an SSE event rather than an HTTP error status:
Validation errors caught before streaming begins are also delivered as SSE events with the code VALIDATION_ERROR:
Your client should listen for events where type is "error" and handle them appropriately — for example, by displaying the message to the user and closing the connection.

​
Rate Limits

thig.ai uses a dual rate limiting system to protect the platform and ensure fair usage:
  • Tier-based limits — Based on your subscription plan (Free, Pro, Business, Enterprise). These govern monthly AI token usage, project counts, and other resource quotas. See Billing for plan details.
  • Endpoint-based limits — Anti-abuse limits applied per endpoint and per IP address. These prevent excessive request volumes regardless of plan tier.
When you exceed a rate limit, the API returns a 429 status with the RATE_LIMITED error code.
Check the Retry-After header in 429 responses. It indicates how many seconds to wait before retrying. Implement exponential backoff in your integration to handle rate limits gracefully.

​
Troubleshooting Tips

Common issues and how to resolve them:
  • “401 Unauthorized” on every request — Your session has likely expired. Log out and log back in, or generate a fresh API key from Settings > API Keys.
  • “403 Feature disabled” — The feature you are trying to use is not enabled for your current plan. Check your plan limits and upgrade if needed.
  • “429 Rate limited” — You are sending too many requests. Implement exponential backoff and respect the Retry-After header. Consider batching requests where possible.
  • “500 on AI endpoints” — The AI provider may be experiencing issues. Try switching to a different provider in your project settings or BYOK configuration.
  • “Empty response from AI” — The conversation or PRD content may be too short for the AI to analyze meaningfully. Add more context to your conversation before generating.
  • “Export fails” — Ensure the project has a generated PRD before attempting to export. If exporting to Notion or Google Docs, verify that the integration is connected.
  • “404 on a resource you just created” — The resource may belong to a different organization context. Verify you are authenticated as a member of the correct organization.
  • “409 Conflict on version creation” — A PRD version with that number already exists. The system auto-increments version numbers, but manual creation may cause conflicts. Use the next available version number.
  • “403 Approval required” — An approval workflow is configured for this action. Your request has been submitted for review. Wait for designated approvers to approve it.
  • “403 IP blocked” — Your IP is not in the organization’s allowlist. Contact your admin to add your IP or CIDR range.
  • “401 API key expired” — Your REST API key has passed its expiration date. Create a new key or extend the expiration from Settings > Developer.