Sprint/v1.2.0#3
Merged
aaddrriann merged 3 commits intomasterfrom Apr 28, 2026
Merged
Conversation
Sprint alignment: - Add scripts/sprint-start.sh — fetches origin/master and compares commit counts using --count to avoid Windows line-ending issues; exits 1 on divergence so callers can branch on exit code. - Add .claude/skills/sprint-align.md (/sprint-align) — protocol for running the script, interpreting output, and guiding rebase/merge decisions without executing them autonomously. Branch lifecycle skills: - Add .claude/skills/new-branch.md (/new-branch) — enforces correct base branch per type: feature/* off dev, fix/* off master, sprint/* and release/* off master and dev respectively. - Add .claude/skills/merge-flow.md (/merge-flow) — validates PR target, blocks direct-to-master routing for non-hotfix branches, and runs a pre-PR checklist covering alignment, version bump, and changelog. - Add .claude/skills/commit-msg.md (/commit-msg) — drafts structured commit messages: section labels, em-dash rationale, 72-char wrap, no Co-Authored-By, no markdown in body. Settings and gitignore: - Add .gitattributes — forces LF on *.sh and .claude/skills/*.md to prevent CRLF corruption of bash scripts on Windows checkout. - Add .claude/settings.json — registers .claude/skills/ as the skill discovery path for Claude Code. - Ignore run.bat and setup.bat — local developer convenience scripts with no value in version history.
Type system & storage: - Add LocalConnection interface and extend MindProvider to include 'ollama' — routes inference to any OpenAI-compatible local endpoint alongside Anthropic and OpenAI. - Extend FlickySettings with localConnections array; add IPC channels for CRUD, bearer key management, and model pull/delete/create. Ollama API service (new ollama-api.ts): - testConnection tries /api/tags then falls back to /v1/models — covers native Ollama and external OpenAI-compat providers (xAI, LM Studio). - normalizeBase strips trailing /v1 to prevent URL path doubling when callers append /v1/chat/completions. - streamChat carries no timeout and no token cap — model runs to EOS; PTT AbortSignal is the sole cancellation mechanism, preventing silent hangs on slow CPU inference (e.g. qwen3-vl:8b at ~8 t/s). - isVisionModel covers 16 known vision families for screenshot routing. IPC & preload: - Eight new handlers in index.ts for connection CRUD, bearer key ops, and Ollama model pull/delete/create with progress event forwarding. - Preload exposes matching methods on window.flicky including streaming pull-progress, pull-complete, and pull-error event subscriptions. UI — MindTab, OllamaSection, ConnectionRow, AddConnectionModal: - MindTab gains a third "Local" tab; selecting it renders OllamaSection which lists enabled connections with per-row model selector. - AddConnectionModal: External/Local type toggle, URL verify with latency + model count hint, bearer auth, prefix ID, model IDs, tags. - OllamaManageModal: installed model list, 11-entry vision model catalog with quick-pull, manual pull-by-tag, delete, and create. - vision-models.ts catalogs 11 vision-capable models with Ollama library slugs for one-click pull and Browse links. Polish: - Manage modal widened to 680px; catalog wrapped in 300px scrollable container; thin 5px pill scrollbar applied to catalog, installed model list, and modal body — consistent with app scrollbar style. - Redundant vision badge removed from catalog cards; section title already reads "Vision-capable models".
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
OllamaManageModal:
- onClick={handlePull} → onClick={() => handlePull()} — prevents
MouseEvent from being passed as the optional tag parameter,
fixing TS2322 that was failing the CI typecheck step.
Release:
- package.json version bumped 1.0.3 → 1.2.0 to align with
sprint/v1.2.0 Ollama local connections feature.
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.
Summary
inference to any Ollama or OpenAI-compatible local endpoint (LM Studio,
vLLM, xAI, etc.)
OllamaAPIservice handles connection testing, model listing,pull/delete/create, and streaming chat — with URL normalization to fix
external provider path doubling
AddConnectionModal,ConnectionRow,OllamaSection, andOllamaManageModalwith an 11-entry vision modelcatalog
AbortSignalisthe sole cancellation mechanism — prevents silent hangs on CPU-only
inference
list both scroll with a consistent thin 5px pill scrollbar
Test plan
list
npm run typecheck— zero errors