Learn / Use Netra Runtime with any OpenAI-compatible tool
Use Netra Runtime with any OpenAI-compatible tool
Netra Runtime serves models through an OpenAI-compatible API, and that one design decision means the entire OpenAI ecosystem is your integration surface. Agent frameworks, workflow builders, chat UIs, coding assistants: if a tool can talk to OpenAI, it can talk to a model you own on Netra. Every integration comes down to the same three values.
The pattern: three values
| Value | What it is | Where to find it |
|---|---|---|
| Base URL | https://api.netraruntime.com/v1 | Same for every account; self-hosted deployments use their own endpoint URL |
| API key | The bearer token that authenticates requests | Netra dashboard |
| Model id | The exact id of the model to serve, e.g. qwen3.6-35b | The /v1/models endpoint or your dashboard |
Before configuring any tool, prove the endpoint works with one request. If this returns a completion, every remaining problem is tool configuration, not the endpoint:
curl https://api.netraruntime.com/v1/chat/completions \
-H "Authorization: Bearer $NETRA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "qwen3.6-35b",
"messages": [{ "role": "user", "content": "Hello" }]
}'Step-by-step guides
| Tool | What it is | Guide |
|---|---|---|
| OpenClaw | Open-source personal AI assistant | Setup guide |
| n8n | Workflow automation with AI nodes | Setup guide |
| Dify | Visual LLM app platform | Setup guide |
| Hermes Agent | Nous Research's autonomous agent | Setup guide |
| LangChain | LLM application framework | Setup guide |
| Open WebUI | Self-hosted chat interface | Setup guide |
| LlamaIndex | RAG and data framework | Setup guide |
| Vercel AI SDK | TypeScript AI toolkit | Setup guide |
| CrewAI | Multi-agent framework | Setup guide |
Other tools that work the same way
These don't have dedicated guides yet, but all of them accept a custom OpenAI-compatible endpoint, so the three-values pattern applies directly:
- Coding assistants: Cline, Continue, and aider all take a custom base URL and model id in their provider settings.
- Chat UIs: LibreChat and LobeChat register custom OpenAI-style endpoints alongside their built-in providers.
- Visual builders: Flowise and Langflow expose base URL fields on their OpenAI chat model nodes.
- Agent frameworks: AutoGen (AG2), Semantic Kernel, and Haystack configure OpenAI-compatible clients with a base URL and key.
- Automation platforms: Zapier and Make can call your endpoint with a plain HTTP POST to
/v1/chat/completions, the same request as the curl test above.
Want a dedicated guide for one of these? Tell us at hello@netraruntime.com.
Common gotchas
- The base URL ends at
/v1. Tools append/chat/completionsthemselves; pasting the full path causes 404s. - Model ids must match exactly. Copy the id from the
/v1/modelslisting rather than retyping it. - Network reachability. A cloud tool needs a publicly reachable endpoint; a VPC or on-prem deployment needs the tool running inside the same network.
- Docker networking. A tool in a container reaching an endpoint on the host must use
host.docker.internal, notlocalhost.
Frequently asked questions
What do I need to connect a tool to Netra Runtime?
Three values: the base URL https://api.netraruntime.com/v1, an API key from your dashboard at app.netraruntime.com, and a model id from the /v1/models endpoint. Any tool with an OpenAI-compatible or custom base URL option accepts them.
My tool isn't listed. Can it still use Netra Runtime?
Almost certainly. Look for a provider option named "OpenAI-compatible", "custom endpoint", or "base URL" in the tool's model settings. If it can talk to OpenAI, it can talk to Netra Runtime with the same three values.
How do I test the connection outside any tool?
Send a curl request to /v1/chat/completions on your base URL with your API key as a Bearer token. If that returns a completion, the endpoint works and any remaining problem is in the tool's configuration.