Add Unstructured Transform MCP cookbook

[dizon-u10d]

Adds a cookbook notebook showing a Haystack Agent calling the hosted Unstructured Transform MCP server to parse and chunk a document end-to-end.

• New: notebooks/unstructured_transform_mcp.ipynb; registered in index.toml (topics: Agents, MCP, Data Processing).
• Uses MCPToolset + StreamableHttpServerInfo's native token param (UNSTRUCTURED_API_KEY), then an AnthropicChatGenerator-backed agent.

Docs: https://docs.unstructured.io/transform/overview · Get started: https://transform.unstructured.io/get-started

🤖 Generated with Claude Code

Companion integration-catalog PR: deepset-ai/haystack-integrations#540

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[dizon-u10d]

Hi @bilgeyucel, following up on this cookbook (Transform MCP via MCPToolset/StreamableHttpServerInfo, using the native token param).

Verified the connection pattern against the live server: reaches https://mcp.transform.unstructured.io correctly and gets a proper auth rejection for an invalid token (see the note on PR #540 about the ~30s timeout behavior with eager_connect=True, same applies here). Registered in index.toml alongside the other MCP entries. Let me know if anything needs adjusting!

[dizon-u10d]

Updated for the recent Transform MCP tool rename: transform_files→start_transform_job, check_transform_status→check_job_status, get_transform_results→get_job_results. Ready for another look whenever you get a chance.

[dizon-u10d]

Small follow-up: corrected the tool count (server exposes 7 tools total now, this cookbook covers the 4 core pipeline ones), added a wait between check_job_status polls, and dropped the OAuth-fallback/mcp-remote note since it wasn't necessary for this integration (the token-auth path shown earlier already works end to end).

[dizon-u10d]

Another round of tool-surface changes shipped upstream this week (file manager tool added), so I went further than a rename this time: rewrote the docs, notebook, and system/agent prompts to describe the pipeline by behavior (submit → poll → fetch) rather than hardcoding exact tool names, and lean on the toolset-discovery pattern already shown here to find the real names at connect time. Should hold up through future renames/additions without needing another update like this.