6c55d80dea
- Real SSE streaming (token-by-token display via Chat Completions API) - Token cost tracking with real DeepSeek pricing in header - Context window usage percentage display - Atomic session writes (crash safety via temp+rename) - Responses API recovery (periodic probe instead of permanent fallback) - Tool parse failure user-visible warnings - Default model changed to deepseek-v3.2 everywhere - V3.2 reasoning/thinking support - V3.2 pricing and context window entries
14 lines
691 B
Markdown
14 lines
691 B
Markdown
# Roadmap
|
|
|
|
This roadmap captures planned work after the 0.3.8 line. Dates are intentionally omitted; versions are the delivery targets.
|
|
|
|
## 0.3.9 (Planned)
|
|
- Duo mode groundwork: refine prompts, validation flow, and UX surfacing for player/coach phases.
|
|
- Documentation: publish a concise Duo spec and user workflow guide.
|
|
- Tooling: tighten phase transitions and reporting in `duo_*` tools.
|
|
|
|
## 0.4.0 (Planned)
|
|
- RLM mode reintroduction: align design with the MIT RLM paper (REPL-style interaction + recursive sub-calls).
|
|
- External memory UX: improved loading/query ergonomics and clear persistence semantics.
|
|
- Documentation: RLM architecture and usage guide, with explicit parity notes.
|