Skip to content

feat: add Auxen as an approved gateway provider#5685

Open
ai-auxen wants to merge 1 commit into
Helicone:mainfrom
auxen-ai:auxen-provider
Open

feat: add Auxen as an approved gateway provider#5685
ai-auxen wants to merge 1 commit into
Helicone:mainfrom
auxen-ai:auxen-provider

Conversation

@ai-auxen
Copy link
Copy Markdown

@ai-auxen ai-auxen commented May 22, 2026

Adds Auxen — per-customer dedicated, OpenAI-compatible LLM endpoints — to Helicone's approved-domain gateway list.

What Auxen is

Each Auxen instance is a dedicated GPU running one open-source model (Llama 3.1/3.2, Qwen 2.5, Mistral, Gemma 2, Mixtral, Phi-3, Command R) on a stable HTTPS URL of the form `https://api.auxen.ai/v1/<instance_id>/v1`, authenticated with a per-instance `auxk_*` bearer token. Billing is per-minute of GPU runtime, not per-token.

Why the dedicated-domain pattern works

The instance ID is path-based, and the API is OpenAI-wire-compatible. Callers just replace the `api.auxen.ai` host with `auxen.helicone.ai` and keep the rest of the path unchanged — no special handling needed in the gateway code beyond the standard `GATEWAY_TARGET` mapping that DeepSeek, Together, etc. already use.

```
https://api.auxen.ai/v1/inst_xxx/v1/chat/completions

https://auxen.helicone.ai/v1/inst_xxx/v1/chat/completions
```

Files touched

  • `worker/wrangler.toml` — register `auxen.helicone.ai` custom domain
  • `worker/src/index.ts` — route `auxen.*` host to `https://api.auxen.ai\`
  • `docs/getting-started/integration-method/auxen.mdx` (new) — integration doc
  • `docs/getting-started/integration-method/gateway.mdx` — add Auxen to approved-domains table
  • `docs/docs.json` — sidebar entry

Note on approval process

Per the gateway docs, dedicated-domain approval is ordinarily a request to engineering@helicone.ai. Filing this as a PR with the proposed wiring to make it easy to review. Happy to follow up with domain ownership verification (TXT record at `api.auxen.ai` or direct contact via `hello@auxen.ai`) if that's the preferred path.

Companion PRs (Auxen distribution series)

Adds Auxen — per-customer dedicated, OpenAI-compatible LLM endpoints
(https://auxen.ai) — to Helicone's approved-domain gateway list.

Each Auxen instance is a dedicated GPU running one open-source model
(Llama, Qwen, Mistral, Gemma, Mixtral, Phi, Command R) on a stable HTTPS
URL of the form https://api.auxen.ai/v1/<instance_id>/v1. Because the
instance ID is path-based and the API is OpenAI-compatible, the gateway
pattern works cleanly: callers replace the api.auxen.ai host with
auxen.helicone.ai and keep the rest of the path unchanged.

Changes:
- worker/wrangler.toml: register auxen.helicone.ai custom domain
- worker/src/index.ts: route auxen.* host to https://api.auxen.ai
- docs/getting-started/integration-method/auxen.mdx: integration doc
- docs/getting-started/integration-method/gateway.mdx: add Auxen to
  approved-domains table
- docs/docs.json: add sidebar entry

Note for maintainers: per the gateway docs, dedicated domain approval is
ordinarily a request to engineering@helicone.ai. Filing this as a PR with
the proposed wiring to make it easy to review; happy to follow up with
domain ownership verification (TXT record at api.auxen.ai or contact via
hello@auxen.ai) if that's the preferred path.

AI agent (Claude) assisted in drafting this PR.
Copy link
Copy Markdown
Contributor

@greptile-apps greptile-apps Bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Your free trial has ended. If you'd like to continue receiving code reviews, you can add a payment method here.

@vercel
Copy link
Copy Markdown

vercel Bot commented May 22, 2026

@Samp2Alex is attempting to deploy a commit to the Helicone Team on Vercel.

A member of the Team first needs to authorize it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants