Providers · Beta

Give each Together AI workload its own limits.

Connect one Together AI project. Give each app an approved chat or embedding model, an immediate spending limit, and a compatible fallback.

Create a free workspace Try the no-card Test Lab

How the connection works

How Together AI works with the gateway

Confirm access without inference

The connection test reads Together's model catalog with the encrypted project key. It does not send a prompt or generate tokens merely to prove the key can reach the project.

Add a limit for each workload

Set a separate gateway key, dollar ceiling, request pace, token pace, and concurrency cap for every app. Together does not offer per-key usage limits; its dynamic organization-and-model limits remain the upstream ceiling.

Keep chat and embeddings distinct

The adapter accepts Chat Completions and embeddings on their documented paths and rejects unsupported protocol boundaries. Together's namespaced model ID must be eligible for the workload's exact request type before it can become a target.

Current shared controls

What every supported connection includes

Encrypted BYOK credentials that remain hidden after setup Stable model aliases and compatible route targets Organization and key budgets, rate limits, and concurrency limits Observe, Shadow, and Enforce modes with metadata-only evidence

Test the exact Together model, request type, streaming behavior, returned usage, and dynamic upstream limit before production approval.

Client boundary

AIGHQ_BASE_URL=https://api.aigatewayhq.com/v1
AIGHQ_API_KEY=from-your-workspace
AIGHQ_MODEL=company-approved-fast

Each Together-backed application uses the gateway URL, its scoped workload key, and a stable model alias. Administrators can change an eligible model or project API key centrally without distributing the Together credential to every app.

Implemented boundary

Supported protocols

Select Together AI and enter a project-scoped API key. The connection check reads Together AI's model catalog without running inference. Then map an eligible namespaced model ID to a stable alias.

POST /v1/chat/completions OpenAI-compatible streaming POST /v1/embeddings GET /v1/models connection test

Compatibility follows the provider's current published interface. Review the provider documentation before approving Together AI for production.