Repository navigation
Chat2DB Community: Start here #1873
Description
Chat2DB Community uses this pinned issue as the single starting point for reporting problems, contributing changes, and joining active design work. Detailed rules live in the version-controlled repository documentation linked below.
Report a problem or request a change
| What you need | Where to go |
|---|---|
| A reproducible Chat2DB problem | Bug report |
| A database-specific connection, metadata, SQL, or editor problem | Database compatibility bug |
| A concrete product improvement | Feature request |
| A documentation correction or addition | Documentation task |
| Installation, configuration, or usage help | Q&A Discussions |
| An early idea or workflow to explore | Ideas Discussions |
| A security vulnerability | Follow SECURITY.md and do not post details publicly |
You can also start from the Issue chooser. Before submitting, search existing Issues and Discussions, use the latest applicable version or provide a source commit, keep one actionable problem per Issue, and remove secrets or private data. See the Issue triage guide for the evidence maintainers need.
Contribute
- Read CONTRIBUTING.md.
- Browse Good First Issues and Help Wanted.
- Claim a published contribution task with
/claimand follow the finding and claiming work contract. - Share verified workflows and integration results in Show and tell.
Open a dedicated Issue or Discussion for concrete work instead of using comments on this navigation Issue.
Active community feature initiatives
Parent Issues organize related Community feature work while keeping each deliverable available as an independently claimable leaf Feature Issue. Open the parent for the complete feature registry, then claim only a linked leaf Issue that is in Project status Ready and has a contribution/* label.
| Initiative | Parent Issue | Focus |
|---|---|---|
| MySQL Community feature support | #2567 - MySQL Community feature support and contribution plan | Complete MySQL product capabilities organized as native, individually claimable Feature sub-issues. |
Active design discussions
Each active design Discussion has one open Tracking Issue:
- The Tracking Issue owns the user problem, scope, priority, acceptance criteria, and implementation tracking.
- The Discussion owns unresolved design questions, alternatives, and decision records.
This list is a navigation index, not a roadmap or delivery commitment. Keep an entry only while its Discussion and Tracking Issue remain active.
Activity
openai0229 commentedon Jul 23, 2026
The maintained contributor entry point now lives in CONTRIBUTING.md, the Community Project, contributor-ready label queries, and .github/COMMUNITY_OPERATIONS.md from #1939. Closing this standalone guide so contribution rules stay in version-controlled repository documentation.
Shubh2-0 commentedon Aug 1, 2026
i'm a backend dev looking to help out with postgres integration and performance. i'd love to take on optimizing schema metadata fetching using a redis caching layer and tuning spring boot connection pool timeouts.
openai0229 commentedon Aug 2, 2026
@Shubh2-0 Thanks for offering to help. Please open a separate issue using the appropriate issue template and clearly describe the current problem, including the existing behavior and impact, reproducible steps and environment, and any baseline or profiling data that shows the bottleneck.
Please also explain the proposed scope and why Redis caching or HikariCP tuning is appropriate for the reported problem. Once the task is clearly defined, you can create a focused pull request and link it to the new issue with
Closes #<issue-number>.Please use the new issue, rather than this contributor guide, to track the implementation.
[-]Contributing to Chat2DB Community: Start here[/-][+]Chat2DB Community: Start here[/+]Shubh2-0 commentedon Aug 6, 2026
i can help out with issue #1942 for handling large ai tool outputs. i'm thinking we can store full result payloads in redis using a configurable ttl and stream preview chunks back to the client via spring webflux to keep memory usage low.
openai0229 commentedon Aug 6, 2026
@Shubh2-0 Thanks for volunteering. Separating a bounded preview from retained result data, enforcing expiration, and avoiding large in-memory payloads are all relevant to #1942.
#1942 is still at the design stage, so we are not ready to commit to a Redis/WebFlux implementation yet. This is a sequencing and architecture-boundary decision, not a rejection of your participation. You are welcome to help shape the design and later take on a focused implementation issue after the accepted work is split.
The current direction is a storage-neutral artifact contract with a local file or embedded implementation first, so Chat2DB Community does not require an external storage service. A Redis-backed implementation can still be evaluated later if it preserves the same AI-tool, authorization, lifecycle, and retrieval contracts.
The design must define streaming materialization, completeness, bounded cursor/limit reads, column projection, filtering capabilities, authorization scope, quota, cancellation, crash recovery, and cleanup. WebFlux transport alone does not resolve those contracts, and storing the current bounded
ExecuteResponsewould not produce a complete result.Please continue the storage-format and retrieval/filtering design in Discussion #1941:
#1941
Issue #1942 remains the tracking Issue for accepted requirements. Once the design is resolved, we will split the work into focused contribution Issues, and you are welcome to take one of those implementation tasks.
If you would prefer implementation-ready work now, you can also browse the Active community feature initiatives in #1873 and the MySQL feature registry in #2567:
#2567
Please claim only a linked leaf Issue that is in Project status
Readyand has acontribution/good-first-issueorcontribution/help-wantedlabel. Parent and Backlog Issues remain planning items rather than directly claimable tasks.Shubh2-0 commentedon Aug 8, 2026
i can help out with the large AI tool result delivery task listed in #1942. i'd implement a redis caching layer with a short TTL for quick previews and stream full artifacts directly into postgres bytea columns.
Shubh2-0 commentedon Aug 11, 2026
i'm interested in picking up backend work for #1942 regarding large ai tool result delivery. i can implement the temporary result caching using redis with configurable ttls and stream the payload chunks via spring boot sse to prevent memory bloat.
Shubh2-0 commentedon Aug 14, 2026
i can help implement the plugin query composition from tracking issue 1911 by creating a unified spi in java and abstracting pagination and metadata query generation across the postgres and mysql dialect layers. let me know if i should go ahead and claim one of the sub-issues to get started.
Shubh2-0 commentedon Aug 15, 2026
i'd like to help with #2566 by implementing the resumable event streams using spring sse emitters backed by redis pub/sub and checkpointing execution state in postgres. let me know if i can claim that task and start putting together a draft pr.
openai0229 commentedon Aug 17, 2026
@Shubh2-0 Thanks for your interest. #1911 is not contributor-ready yet because the request contract and plugin ownership boundaries are still being designed in Discussion #1944. Please share your concrete proposal there first, including the structured filter/sort request shape, the plugin SPI and ownership boundaries, compatibility for MongoDB and relational databases, the migration approach, and the test plan. Once the design is accepted, we can split the work into scoped sub-issues for implementation.
Shubh2-0 commentedon Aug 19, 2026
i can help implement the artifact retrieval flow for large tool results in #1942 by setting up a local file streaming pipeline backed by redis ttl eviction for cached query outputs. let me know if you'd like me to take this on and i'll open a pr.
Shubh2-0 commentedon Aug 24, 2026
i'd like to help implement the backend agent runtime work in #2566, particularly around building resumable sse event streams and task lifecycle state persistence in spring boot. i can also set up redis caching with custom ttl eviction to handle local artifact storage for the large result preview flow in #1942.