Conversation
- Add memory management endpoints for authenticated users to create, list, delete, and clear memories - Enforce unique `(user_id, content)` memory rows with a migration and model constraint - Switch chat and voice memory capture from marker parsing to native LLM tool calls - Make memories global across learning languages and preserve them when deleting a language - Update memory schemas, native language validation, and prompt/context handling - Refresh project docs, changelog, README, and test counts for the new release
- Change the 1.8.28 entry date in CHANGELOG.md from 2026-08-02 to 2026-08-10
- Add provider tool-compatibility detection and retry LLM turns without tools when unsupported - Limit native memory execution to one call and keep continuation fallback silent on failure - Move chat LLM usage persistence to a separate session and emit memory updates only after confirmed saves - Track voice-session tool capability so unsupported tools are disabled only for the current WebSocket session - Tighten SSE parsing to reject truncated final events and require terminal done or error events - Update chat UI handling for SSE completion, error, and memory update events - Add tests for unsupported-tool fallback, continuation fallback, one-call execution, session-local voice capability, FIFO memory eviction, and truncated SSE detection - Refresh documentation and skill metadata with revised backend and frontend test totals and memory behavior
- Stream visible LLM text immediately during native tool calls and continuation - Prevent fallback retries after partial output has been emitted - Add per-user row locks to single-memory deletion and clear-all operations - Update LLM and memory service tests for progressive streaming and locking - Refresh architecture, testing, and changelog docs to match the new behavior - Adjust recorded backend coverage figures in documentation and skill metadata
- Add `GET /api/study-plan/lessons` to expose generated lesson metadata and completion state - Make lesson completion idempotent by returning completed lessons unchanged - Pass lesson state through the plan page and unit drawer to show Start, Continue, or Review actions - Update lesson pages to block edits in review mode and return to the plan - Add lesson action translations and bump app/version metadata to v1.8.30 - Update changelog, AGENTS notes, API specs, and frontend/backend tests for the new flow
- Lock lesson rows during completion to serialize concurrent requests - Commit lesson completion, progress, XP, and competencies atomically - Move freemium usage recording after a successful database commit - Bypass quota checks for already-completed lesson retries - Force-refresh frontend freemium status after successful completion - Disable duplicate lesson completion submissions in the UI - Add backend and frontend tests for rollback, retries, and freemium caching - Update docs, specs, and test counts to reflect the new behavior
- Adjust version 1.8.30 entry date from 2026-08-04 to 2026-08-05
- Align multiline template literal expressions in the app layout - Preserve existing navigation and trial-day logic while cleaning indentation - Normalize conditional className formatting in desktop, admin, and mobile links
feature: fixes in memory function, add review lesson button in my plan.
- Add OpenAI GPT-5.6 `reasoning_effort="none"` for native tool requests and continuations - Retry tool-enabled requests without tools on pre-output incompatibility or stream failure - Keep partial responses on post-output tool continuation failures instead of duplicating output - Log one-time native tool availability and fallback reasons at INFO with provider and model - Expand llm adapter tests for fallback paths, reasoning effort, and availability logging - Update changelog and architecture/testing docs for new behavior, test count, coverage, and timeout values
- Add explicit fallback prompts without memory-tool instructions for chat and conversation turns - Emit memory-updated events from tool result events and preserve them through cancellation - Reset partial streamed text when tools fail after output and retry with no-tools fallback - Treat empty fallback responses as errors instead of successful assistant messages - Improve tool-incompatibility detection across wrapped errors and common provider messages - Ignore memory-read failures in chat prompt assembly and update prompt-building helpers - Refresh changelog, agent notes, and test documentation for the new streaming behavior - Expand backend tests for tool resets, fallback prompts, cancellation handling, and usage tracking
feature: fixes
- Reduce default public pricing copy in dev, example, and compose envs - Rename the pending-lesson action from Continue to Resume across locales - Add a localized lesson-count heading to the unit drawer - Update unit drawer styling for wider desktop layout, primary actions, and sticky close footer - Adjust frontend tests and architecture docs to match the new drawer behavior
- Raise displayed monthly price from 9.95 to 14.95 - Raise displayed yearly price from 99.50 to 149.50 - Sync new pricing defaults in .env.example, .env.dev, and both compose files
feature: design for detail unit plan
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.