Keep provider keys in your environment.
Provider credentials can stay on your laptop, server, Tailnet host, or private router. A TEKIZ.AI account adds its own API key, usage limits, activity records, and billing.
Compare model access options
OpenRouter gives you hosted models through one API. TEKIZ.AI connects provider accounts, local models, and private servers your team approves, then follows the rules you set for each request.
Use only provider accounts and subscriptions you are allowed to connect. OpenRouter can remain one of your TEKIZ.AI providers; OpenRouter usage is billed separately unless your TEKIZ.AI plan explicitly says otherwise.
Provider credentials can stay on your laptop, server, Tailnet host, or private router. A TEKIZ.AI account adds its own API key, usage limits, activity records, and billing.
OpenRouter is built for broad remote model access through one hosted account. TEKIZ.AI helps a team combine approved provider accounts, local models, and private servers in one setup.
Add OpenRouter alongside local Ollama, OpenAI, Anthropic, Gemini, NVIDIA NIM, and private OpenAI-compatible services. Your team decides which services can handle each type of work.
| Buying question | TEKIZ.AI | OpenRouter |
|---|---|---|
| What are you buying? | One setup for approved models. Hosted keys, usage limits, request history, service order, local or Tailnet deployment, and support around customer-authorized model access. | Hosted model marketplace/API. A broad model catalog through one hosted OpenAI-compatible service. |
| Where do provider keys live? | Local-first by default. Keys can stay on customer-controlled routers or hosts; hosted accounts manage generated TEKIZ.AI Router keys. | Hosted account configuration, including documented BYOK paths and provider fallback controls. |
| Does it route local models? | Yes. Ollama, Ollama Cloud through local Ollama, Tailnet routers, and private OpenAI-compatible endpoints can sit beside cloud providers. | Primarily remote hosted model endpoints and provider routing through the OpenRouter service. |
| Can agents use one base URL? |
Yes. Use https://api.tekiz.ai/v1, local
http://127.0.0.1:8790/v1, or a Tailnet router
endpoint.
|
Yes. OpenRouter exposes an OpenAI-compatible API endpoint for supported clients. |
| How does fallback work? | Your team sets the service order using model capability, provider health, response time, and its own rules. | A hosted model catalog with documented provider-key and service-order options. |
| How do you avoid lock-in? | Run the open-source router locally, point agents at one endpoint, and keep authorized provider access portable. | Use the hosted OpenRouter API and account model; BYOK can reduce provider-custody lock-in for configured providers. |
| Best fit | Production agents, local-first teams, private infrastructure, and teams that need clear rules for model use. | Fast access to many remote models through one hosted marketplace-style API. |
OpenRouter behavior changes over time; confirm current BYOK, fallback, privacy, and provider-policy details in OpenRouter’s own documentation before making a procurement decision.
Migration path
Most clients need only a new base URL, API key, and model name.
Start with tekizai/frontier. On Pro and Max,
tekizai/fusion can run independent tasks in
parallel and review the combined result.
export OPENAI_BASE_URL="https://api.tekiz.ai/v1"
export OPENAI_API_KEY="sk_tekiz_..."
curl "$OPENAI_BASE_URL/chat/completions" \
-H "Authorization: Bearer $OPENAI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "tekizai/frontier",
"messages": [{"role": "user", "content": "Route this safely"}]
}'
sk_tekiz_* key.
Create API key next
Use both services
If you already use OpenRouter, add it as one route in TEKIZ.AI. Your tools use one address while TEKIZ.AI applies the choices your team has made for local models, direct providers, private servers, and OpenRouter.
Private-beta demand
Managed model access is in private beta. Tell us which providers and models you need, and we will confirm what we can include, the usage limits, and the price before anything is enabled.
We save your company contact, providers, preferred billing setup, support needs, and target date. Never send prompts, credentials, API keys, tokens, provider responses, costs, or customer data.