Providers · Beta

Give each DeepSeek workload its own limits.

Connect one DeepSeek API account. Give each app an approved chat model, an immediate spending limit, rate ceilings, and a compatible fallback.

Create a free workspace Try the no-card Test Lab

How the connection works

How DeepSeek works with the gateway

Keep the current protocol boundary explicit

Applications call /v1/chat/completions while the adapter uses DeepSeek's documented unversioned API base. The current adapter rejects Responses, embeddings, and Messages rather than silently sending them to an unverified path.

Separate workloads below the account ceiling

Every app receives a scoped gateway key, dollar ceiling, request pace, token pace, and concurrency cap. DeepSeek still combines traffic across API keys when it calculates account-level concurrency, so that upstream ceiling remains in force.

Settle from returned usage

DeepSeek returns prompt, completion, cache-hit, cache-miss, and reasoning-token detail. Gateway receipts retain normalized usage and cost evidence without rounding a low-cost request down to zero; the DeepSeek account balance remains the upstream hard stop.

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 DeepSeek model, Chat Completions payload, streaming keep-alives, returned usage, account concurrency, and fallback 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 DeepSeek-backed application uses the gateway URL, its scoped workload key, and a stable model alias. Administrators can change an approved chat model, DeepSeek API key, or compatible fallback centrally. The current gateway adapter accepts Chat Completions only.

Implemented boundary

Supported protocols

Select DeepSeek and enter an API key. The connection check reads the model catalog without running inference. Then map an approved chat model to a stable alias; applications keep the gateway's /v1 client path while the adapter calls DeepSeek's unversioned API base.

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

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