Daishi Studio

ai.daishiv1.0.0Updated Sep 30, 2026

Run AI agent evaluations on your own model keys, and publish runs others can review and re-run.

VerifiedStreamable HTTPWeb executableAI & ML

Overview

AI-generated overview

Lets an assistant run AI agent evaluations using your own model keys and publish the runs for others to review and re-run.

What it does
Daishi Studio is a remote MCP endpoint that lets an assistant run AI agent evaluations against your own model keys. It also supports publishing those runs so other people can review them and re-run them. The manifest lists no tools, so the exact operations are not documented here.
When to use it
Consider it when you want an assistant to drive evaluation runs for AI agents and share reproducible results with reviewers or teammates. It is less relevant if you only need local model calls or have no evaluation workflow.
Requirements
A remote MCP client that can connect over streamable HTTP to the provider's endpoint. The manifest declares no authentication, environment variables, or headers, but you will need your own model API keys for the evaluations it runs.
Before you install
The manifest declares no authentication, so verify how access to your runs is controlled before connecting. Evaluations consume your own model keys and may incur provider costs. Publishing runs may expose prompts, outputs, or data to others; review what is shared before publishing.

Installation

In SourceWeft

  1. Open Daishi Studio in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Other MCP clients

Add this to your client's mcpServers config.

{
  "mcpServers": {
    "studio": {
      "type": "http",
      "url": "https://daishi.ai/mcp/studio"
    }
  }
}

Tools

0
Tool metadata has not been indexed yet.

Version history

1
  1. v1.0.0LatestSep 30, 2026