Skip to content

Msty

Add the gateway to Msty Studio as a model provider and every chat with its models is checked against your policies, charged to your team and recorded in Request logs. Split chats can then compare GPT, Claude and Gemini side by side, all on governed calls.

What it is

Msty Studio is a chat app for desktop and the web that runs local models and online providers side by side. Each person runs their own copy. An OpenAI-compatible provider points it at any server that speaks the OpenAI format.

The two settings

Open Settings › Model Providers, choose Add Provider and pick the OpenAI-compatible type:

Field Value
Endpoint (base URL) https://<your-gateway>/v1
API key Your Vidai key (starts vk_), never a provider's key
Models Fetch Models to list the models your key may call, or Add Custom Model with a name
Name Optional, for example "Vidai"

Choose Save. Providers belong to a Msty workspace; the provider form can copy the connection to your other workspaces.

The gateway's address is under Connect your app on the Keys page and in System › Settings › Gateway. You get your key from Keys › Issue a key; its secret is shown once.

A minimal call

  1. Save the provider.
  2. Open a new conversation and pick one of its models, for example claude-sonnet-4.5.
  3. Send "Reply with exactly: ok". A short answer means the address and key are right.

Streaming

Msty streams answers. A stream from Claude or Gemini arrives as OpenAI chunks; Vidai records the tokens and the cost either way.

Tool calls

Msty's tools (MCP servers) travel as OpenAI tool calls and work on every provider the gateway translates to. Each round trip is its own call in Request logs.

Reading what the gateway did

Msty doesn't show response headers. Look the call up in Request logs: filter on Key, Model or Path (/v1/chat/completions). Cost is never in the answer.

Errors

Msty shows the gateway's error message in the conversation. Common ones: 401 invalid_api_key, 403 model_not_allowed, 403 content_guardrail_blocked (a guardrail stopped the message), 404 on every call (the address is missing /v1), 429 (a limit on your key or team). The full list is in Headers and errors.

Model names

Fetch Models calls GET /v1/models, which the gateway answers with exactly the models your key may call, named as your organisation approved them under Models in the Control room.

What goes through the gateway

Msty feature Through the gateway
Chats, split chats (each pane with a gateway model), images in messages, tools Yes, to OpenAI, Azure OpenAI, Anthropic, Gemini and Vertex AI models
Knowledge stacks' embeddings, with an embedding model from the gateway provider OpenAI and Azure OpenAI embedding models only
Local models and other providers with their own keys No: a split chat pane on a local model stays on your machine

Check it works in a minute

Send "Reply with exactly: ok" to a gateway model, then open Request logs a few seconds later and filter on your key.

Good to know

  • Msty is one person's app, so each person uses their own key, issued in Keys › Issue a key; their calls show under them and are charged to their team.
  • Online providers, the OpenAI-compatible one included, are on Msty's free plan; only some provider types (such as Azure and Bedrock) need a paid plan.
  • Msty Studio on the web calls providers from the browser, and the gateway doesn't answer a browser's cross-origin (CORS) check: use the desktop app.
  • Msty's own docs: Online providers.