Starter prompts for the most common discussion types. Copy the relevant block when you open a new discussion — filling in the headings is the fastest way to get a useful reply.
These are also the bodies you'd port into .github/DISCUSSION_TEMPLATE/<slug>.yml form templates if you adopt them later.
For setup help, usage questions, troubleshooting, deployment help.
### What are you trying to do?
<!-- The actual goal, not just the symptom -->
### What did you expect to happen?
### What happened instead?
<!-- Errors, unexpected output, missing behavior -->
### Environment
- Statewave version:
- Install method: <!-- pip / docker / source -->
- Python / Node version:
- Database / storage: <!-- Postgres version, pgvector version -->
- Compiler mode: <!-- heuristic / llm -->
- Model / provider: <!-- via LiteLLM, e.g. gpt-4o-mini, claude-3-haiku, ollama/llama3 -->
- Deployment style: <!-- local / self-hosted / cloud -->
### Relevant logs or code
<!-- Trim to the failure. Redact secrets. Use ``` fences. -->For product / SDK / API ideas, integrations, admin UI improvements. Use the Ideas & Feature Requests category.
### Problem
<!-- What's hard or impossible today? -->
### Proposed solution
<!-- What you'd like to see. Sketch the API or UX if you can. -->
### Why this matters
<!-- Who is blocked / slowed down, and how often -->
### Alternatives considered
<!-- Workarounds you tried, related tools, why they fall short -->
### Who would use this?
<!-- Yourself, your team, a class of users -->
### Urgency
<!-- Blocking, important-not-urgent, nice-to-have -->For design proposals significant enough to deserve community review before implementation. Use the RFCs category.
### Summary
<!-- One paragraph: what is being proposed and why -->
### Motivation
<!-- The problem being solved, with concrete use cases -->
### Proposed design
<!-- Enough detail that someone could critique or implement it.
Cover: data model, API surface, behavior, migration path. -->
### Alternatives
<!-- Other designs considered and why they were rejected -->
### Risks / tradeoffs
<!-- What this constrains, what it makes harder, what could go wrong -->
### Open questions
<!-- Specific things you want feedback on -->
### Feedback requested
<!-- Who you most want to hear from, by when if there's a deadline -->For agents, integrations, memory workflows, evals, benchmarks built with Statewave.
### What did you build?
<!-- Short description, link if public -->
### What does Statewave remember?
<!-- Subjects, episode kinds, what gets compiled, what retrieval looks like -->
### Stack
- Agent framework:
- LLM provider / model:
- Statewave compiler mode:
- Embedding provider:
- Deployment:
### Demo / screenshots
<!-- Optional but appreciated -->
### What worked well?
### What was missing?
<!-- Honest gaps help us prioritize -->For sharing how Statewave plugs into a specific framework or stack. Use the Integrations category.
### Integration target
<!-- e.g. LangChain, LlamaIndex, CrewAI, AutoGen, OpenAI Agents SDK, MCP, custom -->
### Use case
### How Statewave fits
<!-- Where in the agent loop / app it sits, what it stores, what it retrieves -->
### Code sketch
<!-- Minimal example. Public repo link is even better. -->
### Friction encountered
### What would make this easier?
<!-- SDK helpers, docs, examples, API additions -->For long-term memory, temporal memory, agent state, retrieval, evals, memory quality, context engineering.
### Question or hypothesis
### Why it matters for memory systems
### What you've read / tried
<!-- Papers, prior art, experiments -->
### What you're proposing to test or discuss
### How Statewave could help (or not)
<!-- Honest framing — research questions don't need to map to product features -->