[yashbudhia]

Description

Adds a Dreamtune MCP cookbook for comparing GPU catalog options and preparing a workload plan in Continue Agent mode. It includes hosted Streamable HTTP with the required custom API-key header, a pinned local stdio alternative, local secret setup, tool policies, troubleshooting, and navigation links.

I work on Dreamtune. This documentation was prepared with AI assistance and reviewed against both projects' source and public documentation.

Checklist

• I've read the contributing guide.
• The relevant docs have been added and linked.
• The configurations and MCP discovery have been verified; no runtime code changed.

Tests

• Both YAML examples parse with @continuedev/config-yaml 1.23.0's native configuration schema.
• Continue's current source MCPConnection connects using the documented stdio settings and discovers all 34 tools from the published @dreamtune/sdk@0.1.2, including compute_catalog, list_tools, and inspect_tool.
• Mintlify's native docs configuration schema, cookbook navigation, MDX compilation, internal route targets, JSON formatting, staged diff checks, and Gitleaks pass.
• The repository's npm run build fails before loading documentation: locked @mintlify/previewing imports a default export from tar, while the repository overrides tar to a version without that export. Running the installed Mintlify binary with --help outside the docs directory reproduces the same error. I have left that dependency issue outside this documentation change.

The connection check used a synthetic placeholder key and performed MCP initialization and tool discovery only. Authenticated catalog execution and an LLM-driven conversation were not run; no paid model/tool execution or compute was launched.

Screen recording or screenshot

Documentation-only cookbook; the MDX compiles and the configuration examples were checked as described above.

[github-actions]

Thank you for your submission, we really appreciate it. Like many open-source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution. You can sign the CLA by just posting a Pull Request Comment same as the below format.

---

I have read the CLA Document and I hereby sign the CLA

---

_{You can retrigger this bot by commenting recheck in this Pull Request. Posted by the CLA Assistant Lite bot.}