AI Provider Directory | 32 AI APIs Documented | AI Error Hub
32 providers monitored · Updated weekly

Every AI API, Every Error, One Place

The complete directory of AI providers we document — from foundation model APIs like Anthropic Claude and OpenAI to specialized services for music, video, and images. Each provider hub includes documented errors, working fixes, live status, and provider-specific quirks.

32Providers
4,247Errors Documented
12K+Developers Served
7-dayUpdate Cadence
FreeForever

Ecosystem — Protocols & Framework

P2 · Growing Coverage

Not AI providers themselves, but critical protocols and orchestration layers connecting them.

No providers match ""

Try a different search term or clear filters. Or request a provider we haven't covered yet.

Browse by Error Type Instead

Not sure which provider you're using? Every error also belongs to a category. Browse fixes by what went wrong instead of who made it.

Frequently Asked Questions

We use six criteria to select providers: developer traction (1,000+ monthly active developers), active community, public documentation, stable API surface, distinct error patterns, and meaningful search demand. See our Methodology page for the full framework.

Larger, more mature APIs like Anthropic Claude and OpenAI have more surface area, more edge cases, and more historical error variants. Newer providers like Kling or Pika are simpler APIs with fewer distinct errors. Our coverage scales with what genuinely exists — we don't pad numbers.

P1 providers (Claude, OpenAI, Gemini) are retested every 30 days. P2 providers every 45 days. P3 providers every 60 days. Any error we get a correction report on gets retested within 48 hours. See our Editorial Standards for the full retesting schedule.

Yes — email info@aierrorhub.com with subject "New Provider Request." Include the provider name, why they matter to your work, and any specific errors you've hit. We add providers based on demand and quality of documentation available.

Status checks run every 60 seconds. Our status dots reflect the most recent check. For production incident response, always cross-reference with the provider's official status page — our monitoring is a convenience aggregator, not authoritative. See our Disclaimer for full caveats.

Not currently — our focus is hosted AI APIs where developers hit provider-specific errors. Self-hosted setups fail in ways that depend more on your hardware and configuration than provider behavior. We may add limited self-hosted coverage in the future based on demand.

Both. Every fix is initially tested in an isolated sandbox to confirm it resolves the error. High-priority fixes (P1 provider errors) are then validated in a production-like environment. Each error page shows the "Last Tested Date" and "Tested By" fields for transparency.

Provider docs describe how the API is supposed to work. We document how it actually breaks in real production usage — with working fixes, root cause analysis, and cross-provider comparison. Think of us as the "post-mortem" resource to complement the "manual."

Can't find what you're looking for?

Missing a provider, spotted a documentation gap, or want to contribute a fix you've verified? We add providers based on developer demand.