#8232: Nebius token factory Provider support
docs
cli
commands
agents
size: M
Cluster:
Provider Support Enhancements
# Add Nebius TokenFactory provider
## What/Why
- Add Nebius as an OpenAI-compatible provider with auth choice, onboarding flows, and implicit provider auto-discovery via /models.
- Add docs page for Nebius Token Fac tory and link it from provider indexes.
- Nebius now appears in provider lists and can auto-load models when NEBIUS_API_KEY is set.
## Testing
test
<img width="912" height="218" alt="image" src="https://github.com/user-attachments/assets/cd8d528f-2556-4975-8779-de8af5928368" />
lint
<img width="602" height="130" alt="image" src="https://github.com/user-attachments/assets/e527fc5e-e830-4df2-b666-4d7048cf3040" />
<!-- greptile_comment -->
<h2>Greptile Overview</h2>
<h3>Greptile Summary</h3>
This PR adds Nebius Token Factory as an OpenAI-compatible provider across docs and onboarding flows. It introduces a new `nebius-token-factory-api-key` auth choice, stores credentials in an auth profile, applies provider config + default model wiring, and adds implicit provider auto-discovery by calling `GET /models` when a Nebius key is present.
Key integration points are:
- Env-key resolution (`NEBIUS_TOKEN_FACTORY` preferred, `NEBIUS_API_KEY` fallback)
- Interactive + non-interactive onboarding support
- Implicit provider creation in `resolveImplicitProviders` with model discovery/default fallback
Main issues to address are consistency between interactive vs non-interactive env var detection, and ensuring that the “apiKey” value stored in generated config (often a source/reference string) is never treated as a literal credential by runtime call sites.
<h3>Confidence Score: 4/5</h3>
- This PR is mostly safe to merge; main risk is a config/env-var inconsistency that can break non-interactive onboarding in some environments.
- Changes are additive and scoped to a new provider. The notable correctness concern is that non-interactive onboarding only checks `NEBIUS_TOKEN_FACTORY` while other paths accept `NEBIUS_API_KEY` as fallback, which can cause surprising failures. A secondary concern is making sure placeholder apiKey references aren’t used as real credentials by downstream code.
- src/commands/onboard-non-interactive/local/auth-choice.ts, src/agents/models-config.providers.ts
<!-- greptile_other_comments_section -->
<!-- /greptile_comment -->
Most Similar PRs
#15991: feat: add Novita AI provider support with dynamic model discovery
by Alex-wuhu · 2026-02-14
79.2%
#7798: feat: add ERNIE provider support
by mattheliu · 2026-02-03
77.4%
#7418: feat (amazon-nova): add Amazon Nova 1P provider
by 5herlocked · 2026-02-02
76.6%
#13295: feat: add Eternal AI provider integration
by peterparkernho · 2026-02-10
75.4%
#2429: added cerebras as a model provider.
by kkkamur07 · 2026-01-26
75.1%
#7272: Models: add SiliconFlow provider
by qychen2001 · 2026-02-02
73.4%
#7051: Add io-intelligence model inference provider
by rajagurunath · 2026-02-02
72.9%
#6484: docs: add LiteLLM + Nebius integration guide
by demianarc · 2026-02-01
72.7%
#8216: Feat: Add sambanova provider
by luisfucros · 2026-02-03
72.7%
#1943: Add jan.ai provider documentation
by csa1234 · 2026-01-25
72.4%