deepset-ai / deepset-ai/haystack-core-integrations
Feature request: discuss optional ClawMem memory backend / integration
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 203
- Forks
- 332
- Avg merge
- 2d 4h
- Merged PRs (30d)
- 80
Description
Summary
I'd like to discuss whether Haystack would be open to an optional ClawMem integration for durable memory in agent and pipeline workflows.
This is not a request to replace Haystack's architecture. The narrower question is whether ClawMem could fit as an optional component/integration for teams that want auditable long-term memory.
Why this seems worth discussing
Haystack's README explicitly describes the framework as giving developers control over retrieval, routing, memory, and generation, and says users can use built-in components for memory or create their own.
That makes this look like a potentially good fit for:
- an external integration component
- a docs/example integration pattern
- a first-party optional connector if maintainers think it belongs in the ecosystem
What ClawMem would add
ClawMem is oriented around:
- durable memory across sessions
- auditability of memory updates
- repository/workspace-scoped memory boundaries
- optional shared memory across collaborators and agents
Clarification I am looking for
Would maintainers be open to one of these directions?
- external integration component
- docs/example integration
- first-party optional connector
- not a good fit for Haystack
Minimal scope
The smallest useful scope seems to be:
- optional integration only
- preserve existing Haystack defaults
- fit into Haystack's component model instead of introducing a competing orchestration layer
If this sounds reasonable, I can turn it into a concrete implementation brief.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with Haystack's README and its component model to understand how optional memory integrations are currently described and structured. The issue is ready when maintainers select a direction—external component, documentation/example integration, first-party connector, or no fit—and the scope is turned into a concrete implementation brief.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- ai, backend
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100