Add Dewey integration

[lambdabaa]

Dewey — managed document intelligence for Haystack pipelines

This PR adds a listing for dewey-haystack, a Haystack 2.0 integration for Dewey.

Components

• DeweyDocumentStore — DocumentStore protocol implementation backed by a Dewey collection
• DeweyRetriever — @component using Dewey's hybrid semantic + BM25 search
• DeweyResearchComponent — @component that runs Dewey's full agentic research loop and returns a grounded Markdown answer with cited sources

Links

• PyPI: https://pypi.org/project/dewey-haystack/
• GitHub: https://github.com/meetdewey/dewey-haystack
• Docs: https://meetdewey.com/docs#haystack

[kacperlukawski]

@ari-bmgf @lambdabaa Thanks for contributing with the integration.

Ie noticed Dewey is fairly new - could you share a bit about the company, as I couldn't find your LinkedIn page or X account? Even a brief note helps us understand the longevity risk for users who adopt this. Also, is there a point of contact who'd own issues and updates to this integration going forward?

We've been trying to be more thoughtful about which integrations we accept, mostly to make sure we can keep documentation and maintenance sustainable long-term.

[lambdabaa]

Hi @kacperlukawski, thanks for your thoughtful note.

Dewey is an API for extracting structured claims from document collections and answering research questions with cited answers. We're early-stage but actively developed and in production use. I'm Ari Aye, the founder, and I'm the right point of contact for this integration. You can reach me at ari@meetdewey.com.

A few things that speak to longevity: the integration is straightforward (thin wrapper over our REST API), so maintenance burden is low. If Dewey ever changed or sunset, the integration would fail loudly rather than silently, making the dependency easy to detect and remove. We're also happy to commit to being responsive on issues, and you can tag me directly on anything that comes up.