Any OpenAI-compatible provider already works via custom; this issue is the one-line nickname.
Verified 2026-07-12:
- Base URL:
https://api.venice.ai/api/v1 (note the /api/v1 path, not /v1)
- Keyless smoke: POST
{base}/chat/completions returns 401 {"error":"Authentication failed"}
- Docs: https://docs.venice.ai/api-reference/endpoint/chat/completions
- Example model:
venice-uncensored-1-2 (GET /models is keyless if you want to browse)
Acceptance: one entry in src/router/upstream.ts PRESETS with a short path-quirk comment, README provider table row, and the preset covered by the existing tests (the https/no-trailing-slash suites pick it up automatically). See CONTRIBUTING.md for the verification rule.
Any OpenAI-compatible provider already works via
custom; this issue is the one-line nickname.Verified 2026-07-12:
https://api.venice.ai/api/v1(note the/api/v1path, not/v1){base}/chat/completionsreturns 401{"error":"Authentication failed"}venice-uncensored-1-2(GET/modelsis keyless if you want to browse)Acceptance: one entry in
src/router/upstream.tsPRESETS with a short path-quirk comment, README provider table row, and the preset covered by the existing tests (the https/no-trailing-slash suites pick it up automatically). See CONTRIBUTING.md for the verification rule.