Model catalog

One endpoint across authorized model routes.

Use Tekiz profiles, OpenAI-compatible providers, Anthropic-compatible routes, Gemini, Ollama, Ollama Cloud, NVIDIA NIM, OpenRouter BYOK, and private gateways without exposing anonymous model traffic.

Sign in first. Create an API key in your account, then connect an authorized provider or add billing when you are ready to route.

The public catalog is discovery metadata. Live /v1/models and model requests require an active generated sk_sage_* key.

Create API key next
Profiles

Tekiz aliases

sage-router/auto, sage-router/balanced, sage-router/frontier, and sage-router/agentic keep clients stable while the router selects the best healthy backend. sage-router/fusion becomes available when verified synthesis capacity is live.

BYOK

Authorized providers

Connect provider keys, subscriptions, local models, or team-approved endpoints you are authorized to use. Hosted plans sell routing control, not unauthorized model access.

Failover

Private and cloud origins

The public edge can select healthy Tailnet installs or hosted fallback origins, while customer credentials stay in the right custody boundary.

gateway-style discovery

Search public route families.

This explorer reads safe public /model-catalog metadata only. It never calls authenticated /v1/models without a customer key.

Catalog JSON
Loading catalog... Fetching safe discovery metadata from /model-catalog.
Family Example models Access boundary
Tekiz profiles sage-router/auto, sage-router/frontier, sage-router/agentic, sage-router/fusion Stable aliases select across authorized providers and healthy fallback routes. Fusion is a Pro/Max synthesis route when verified capacity is live. See the guide.
OpenAI and Codex-compatible gpt-5.5, gpt-5.4, gpt-5.4-mini Bring an authorized API key or compatible Codex/OpenClaw auth profile. Codex setup supports hosted and local port 8790 profiles.
Anthropic-compatible claude-opus, claude-sonnet, claude-haiku Bring authorized Anthropic/Dario access. Public plans do not pool personal accounts.
Google Gemini gemini-3-pro, gemini-3-flash, gemini-2.5-pro Bring authorized Google API or Vertex credentials.
Ollama and Ollama Cloud llama3.3, qwen3-coder, gpt-oss:cloud Use customer-controlled local Ollama or authorized Ollama Cloud models discovered through the runtime.
OpenAI-compatible BYOK nvidia-nim/nemotron, openrouter/free-models, darkbloom/custom Bring authorized provider keys for compatible endpoints such as NVIDIA NIM, OpenRouter, or private gateways.

Client setup

Keep the model API private.

Use the public catalog to choose a route family, then verify the live model list with your account key:

OPENAI_BASE_URL=https://api.tekiz.ai/v1

OPENAI_API_KEY=sk_sage_...

curl -H "Authorization: Bearer $OPENAI_API_KEY" https://api.tekiz.ai/v1/models