Skip to content
OctaFuse OctaFuse

Providers

In Admin, open Inference → Providers. A provider is where the gateway forwards traffic. For each record you can set OpenAI-compatible, Anthropic, and Gemini endpoints, and maintain one or more upstream API keys (usually masked / fingerprinted after save).

OctaFuse Admin Providers page showing provider stats, search, import, and provider cards.
Providers is where you maintain upstream vendors, protocol endpoints, and key status. Imported presets still need real API keys.

What to do in Admin

Section titled “What to do in Admin”
  1. In the sidebar under Inference, open Providers.
  2. Create or edit: set id, display name, per-protocol endpoints (base or full capability URLs — follow upstream docs for /v1 paths), the first API Key, optional notes.
  3. Multiple keys: add more upstream keys on the same Provider with labels, priority, optional weight, RPM / TPM / concurrency, and status. Scheduling uses key priority / headroom first, then weight on near ties — this is not the same layer as route priority.
  4. Import from catalog: use the built-in catalog to pre-fill endpoints; you still need a valid API Key. For Images, confirm generations (and edits when needed); see Images.
  5. After saving, go to Routing to bind models to this vendor and protocol; use Playground or Simulator to validate connectivity.

Notes

Section titled “Notes”
  • Wrong base URLs / endpoints often show up as 404/401 — double-check upstream docs.
  • This screen does not include a one-click connectivity test; validate via routes + Playground / Simulator.
  • When upstream rotates keys, update them here; you can keep multiple keys for gradual cutover.

Automation

Section titled “Automation”

For bulk changes from your backend, use the admin HTTP API — see the GitHub technical reference.

Next

Section titled “Next”