A privacy-first, AI-powered code editor built for developers who want AI assistance without sending their code to the cloud.
QuillAI runs entirely on your machine when using a local LLM backend. No telemetry. No code uploaded to third-party servers. Your codebase stays yours.
Built with PyQt6. Supports local llama.cpp, any OpenAI-compatible API, and Anthropic Claude.
Every major AI coding tool — Copilot, Cursor, Tabnine — routes your code through external servers. QuillAI doesn't have to. When configured with a local LLM:
- Your code never leaves your machine
- No API keys required
- No usage limits, no subscription
- Works offline
When you do want cloud power (Claude, GPT-4, OpenRouter), you switch with one click in the status bar. The choice is always yours.
Download the latest AppImage from the Releases page:
chmod +x QuillAI-*.AppImage
./QuillAI-*.AppImageNixOS users: AppImages require
appimage-runor system-level support.nix run nixpkgs#appimage-run -- ./QuillAI-*.AppImageOr add to
configuration.nixfor double-click support:programs.appimage = { enable = true; binfmt = true; };
# Run directly
nix run github:GSSparks/quillai
# Or install to your profile
nix profile install github:GSSparks/quillaigit clone https://github.com/GSSparks/quillai
cd quillai
pip install PyQt6 requests pyyaml markdown chardet
python main.pyFor Nix development:
nix develop
python main.py- 🏠 Local (llama.cpp) — FIM completions via a local llama.cpp server. Zero latency, zero cost, zero data sharing. Recommended: Qwen2.5-Coder.
- ☁️ OpenAI / compatible — any OpenAI-style API including OpenRouter, LM Studio, Ollama, and others
- 🟠 Anthropic (Claude) — native Claude API with separate models for chat and inline completions
Switch backends at any time with the mode button in the status bar.
QuillAI features a lightweight auto-discovery plugin system. Drop a new plugin folder into plugins/features/ and it is loaded automatically on next launch — no changes to core code required.
Each plugin is a self-contained package with its own widget code and a thin main.py entry point:
plugins/features/
└── my_feature/
├── __init__.py
├── main.py # FeaturePlugin subclass — activate(), deactivate()
└── my_feature.py # Widget implementation
Plugins communicate via a named event bus (file_opened, file_saved, project_opened, and more — see EVENTS.md). Disabling a plugin is as simple as setting enabled = False in its class body. The following panels are implemented as plugins:
- Terminal — embedded terminal dock
- Import Graph — dependency graph visualization
- Symbol Outline — LSP-powered class/method tree
- Markdown Preview — live preview with scroll sync
Completions are informed by your whole session — recent chat exchanges, pinned memory facts, files you've been editing, and functions you've been working in. Ghost text appears at natural pause points. Tab to accept, Ctrl+Right for word-by-word, Ctrl+Space to trigger manually.
The chat panel understands your entire project: active file and the symbol you're working in, all open tabs, direct and transitive imports (up to 3 levels deep), LSP hover docs and live diagnostics, a structural repo map of the whole codebase, semantic search across your codebase history, and your memory facts. Responses stream live with syntax highlighting and markdown rendering. Code blocks have a one-click copy button.
QuillAI connects to language servers automatically when installed, giving you IDE-grade code intelligence across multiple languages:
- Hover tooltips — signature and docstring for any symbol, shown on mouseover with formatted markdown rendering
- Ctrl+Click go-to-definition — jump to where a function or class is defined, across files
- Diagnostic squiggles — live error and warning underlines as you type
- Breadcrumb bar — always-visible
file › class › methodnavigation at the top of the editor; click to jump - Symbol outline panel — full tree of classes, functions, and variables in the sidebar; click to jump
- LSP completion dropdown — context-aware completions with type signatures and docstrings
- Context-aware chat — LSP hover info and diagnostics are automatically injected into every chat prompt
Supported servers (all included in the Nix package):
| Language | Server |
|---|---|
| Python | python-lsp-server |
| YAML / Ansible | yaml-language-server |
| JavaScript / TypeScript | typescript-language-server |
| Bash / Shell | bash-language-server |
| HTML / CSS / JSON / Markdown | vscode-langservers-extracted |
| Nix | nil |
| Lua | lua-language-server |
| Perl | perlnavigator |
LSP degrades gracefully — everything works normally if a server is not installed.
Split the editor horizontally or vertically to view multiple files side by side. Each pane has its own tab bar and active indicator. Panes collapse automatically when their last tab is closed.
Ctrl+\— split active pane side by sideCtrl+Shift+\— split active pane top/bottomCtrl+Shift+W— close active paneCtrl+K Left/Right— move focus between panes
Visualize your project's import structure as an interactive force-directed graph. Node size reflects connectivity. Double-click any node to open that file. Drag nodes, scroll to zoom, pan by dragging the background. Filter low-connectivity nodes with the min-degree slider.
Supports Python, JavaScript/TypeScript, YAML, Nix, Bash, Lua, and Perl.
A sidebar tree of every class, function, and variable in the current file, powered by LSP documentSymbol. Classes nest their methods. Click any symbol to jump directly to its definition. Updates live as you edit with a 1.5s debounce.
QuillAI builds a structural map of your entire project on open — every file, every class, every function signature and docstring — and injects a query-filtered slice of it into every chat prompt. The model gets a navigational overview of the whole codebase without the token cost of full source. Ansible playbooks and role imports are followed and included.
The map is built in a background thread on project open, invalidated on file save, and filtered per-query so only structurally relevant files are included.
QuillAI builds a semantic search index over your entire project that grows smarter with use. Four collections are indexed and searched on every chat message:
- Code — every function and class, chunked by AST symbol, searchable by meaning not just name
- Conversations — past chat exchanges, so similar problems surface relevant history
- Completions — ghost text you have accepted, building a model of your patterns over time
- Edit patterns — files you edit together, surfaced when working in related areas
Uses sentence-transformers locally (no API required) or OpenAI embeddings when in cloud mode. The longer you use QuillAI on a project, the more relevant its suggestions become.
QuillAI remembers things across sessions:
- Global facts — preferences that apply to all your work
- Project facts — things specific to the current codebase
- Conversation history — past exchanges, searchable, clickable to restore
- Turn buffer — recent messages are always included verbatim so the AI has genuine conversational continuity within a session, not just summaries
Facts are auto-extracted from your chat messages. Everything is stored locally at ~/.config/quillai/.
Full multi-cursor support — every keystroke, deletion, and paste applies to all cursors simultaneously with atomic undo:
Ctrl+D— add cursor at next occurrence of selected word (press again to step through)Ctrl+Shift+L— add cursors at all occurrences in the fileCtrl+Alt+Up/Down— add cursor on line above/below (column mode)Alt+Click— add cursor at any position (click again to remove)Escape— clear all secondary cursors, return to single cursor
QuillAI autosaves every 2 minutes to ~/.config/quillai/autosave/. If the app crashes, your work is silently restored on next launch — no dialog, no friction. Recovered tabs are marked with ↩ in their title until saved. On clean exit, autosave files are removed automatically.
Ctrl+P — fuzzy search across open tabs, all project files, and editor actions in a unified list. Arrow keys or Tab to navigate, Enter to open, Esc to dismiss.
**Ctrl+\``** — toggle a full terminal docked at the bottom. Uses qtermwidget` for a full PTY experience when available, with a QProcess-driven interactive shell as fallback. Working directory follows the open project automatically.
Each project remembers which files you had open and where your cursor was. Switching projects restores that project's tabs, chat history, and memory. Recent Projects menu with tab count for each entry.
- Syntax highlighting for Python, HTML, Ansible/YAML, Nix, Bash, Markdown, Perl, and more
- Line numbers with live git diff indicators (green = added, amber = modified)
- Double-click line number to select the entire line
- Minimap with click-to-navigate and viewport highlight
- Smooth scrolling with ease-out animation
- Git blame in the gutter — toggle per-file to see commit hash and author per line
- Bracket match highlighting
- Indent guides, auto-closing brackets, smart auto-indent
- Color swatch inline for hex color values — click to open color picker
Ctrl+E— AI rewrite of selection with side-by-side diff previewCtrl+I— inline chat popup at the cursorCtrl+G— jump to line,Ctrl+Shift+D— duplicate line,Ctrl+/— toggle comment
Chat and Memory live in a sliding panel on the right edge. Hover to expand, pin to keep open, drag the left edge to resize. Width persists across sessions.
Opens automatically when editing .md files. Live preview with full CommonMark support. Scroll position syncs with the editor cursor — the preview follows as you move through the document. Floatable — drag it wherever you want on screen.
Changed files tree, selective staging with checkboxes, inline diff viewer, commit/push/discard — plus AI-generated commit messages from your staged diff.
Ctrl+F— live find/replace with match countCtrl+Shift+F— search across the entire project
F5 — run the current Python script. Output panel with stdout/stderr and a 💡 Explain Error button that sends the traceback to the AI chat.
Ctrl+Shift+Space — fuzzy search across built-in snippets for Python, Ansible, Nix, and Bash. User-editable at ~/.config/quillai/snippets.json.
llama.cpp:
./server -m your-model.gguf --port 11434 -c 8192Then in QuillAI settings (Ctrl+,):
- Server URL:
http://localhost:11434/v1/chat/completions
Recommended models:
- Chat:
Qwen2.5-Coder-32B-Q4_K_M(32GB VRAM) orQwen2.5-Coder-7B-Q4_K_M(8GB VRAM) - Inline completions: any FIM-capable model, 7B or smaller for low latency
Open File → Settings (Ctrl+,):
| Section | Setting | Description |
|---|---|---|
| Local LLM | Server URL | llama.cpp or compatible endpoint |
| Local LLM | Inline model | Fast model for ghost text |
| Local LLM | Chat model | Larger model for chat |
| OpenAI | API URL | Defaults to api.openai.com |
| OpenAI | API Key | sk-... |
| OpenAI | Chat model | e.g. gpt-4o |
| Anthropic | API Key | sk-ant-... |
| Anthropic | Chat model | e.g. claude-sonnet-4-6 |
| Anthropic | Inline model | e.g. claude-haiku-4-5-20251001 |
All settings stored locally at ~/.config/quillai/settings.json.
| Key | Action |
|---|---|
Ctrl+P |
Command palette |
Ctrl+Space |
Trigger inline AI completion |
Tab |
Accept full ghost text suggestion |
Ctrl+Right |
Accept next word of suggestion |
Ctrl+Shift+Space |
Open snippet palette |
Ctrl+E |
AI rewrite of selection (with diff preview) |
Ctrl+I |
Inline chat at cursor |
Ctrl+Click |
Go to definition (LSP) |
Ctrl+Return |
Send chat message |
| `Ctrl+`` | Toggle terminal |
Ctrl+\ |
Split editor pane (side by side) |
Ctrl+Shift+\ |
Split editor pane (top/bottom) |
Ctrl+Shift+W |
Close active pane |
Ctrl+K Left/Right |
Focus adjacent pane |
Ctrl+D |
Multi-cursor: add next occurrence |
Ctrl+Shift+L |
Multi-cursor: add all occurrences |
Ctrl+Alt+Up/Down |
Multi-cursor: add cursor above/below |
Alt+Click |
Multi-cursor: add cursor at position |
Escape |
Clear secondary cursors |
Ctrl+G |
Go to line |
Ctrl+Shift+D |
Duplicate line or selection |
Ctrl+/ |
Toggle comment |
Ctrl+] |
Indent selection |
Ctrl+[ |
Unindent selection |
Ctrl+F |
Find / replace |
Ctrl+H |
Find / replace (focus replace field) |
Ctrl+Shift+F |
Find in files |
Ctrl+N |
New tab |
Ctrl+Shift+N |
New project |
Ctrl+O |
Open file |
Ctrl+S |
Save |
Ctrl+Shift+S |
Save as |
Ctrl+, |
Settings |
F5 |
Run script |
Python 3.10+
PyQt6
requests
markdown
pygments
Optional but recommended:
pyyaml # YAML/Ansible linting
chardet # Encoding detection
python-lsp-server # LSP for Python
sentence-transformers # Local embeddings for vector index
chromadb # Vector index storage
pyqtermwidget # Full PTY terminal (Linux/macOS)
shellcheck # Bash linting (via system package manager)
perlnavigator # LSP for Perl
quillai/
├── main.py # Main window and application entry point
├── EVENTS.md # Plugin event bus reference
├── ai/
│ ├── worker.py # AIWorker — all LLM backends and streaming
│ ├── context_engine.py # Context assembly — symbols, imports, LSP, repo map, vector
│ ├── lsp_client.py # Generic JSON-RPC LSP client
│ ├── lsp_manager.py # Multi-server LSP registry and routing
│ ├── lsp_context.py # Formats LSP hover/diagnostics for chat context
│ ├── repo_map.py # AST-based structural project map (Python + Ansible)
│ ├── embedder.py # Embedding router (local sentence-transformers / OpenAI)
│ ├── vector_store.py # ChromaDB wrapper, per-project collections
│ └── vector_index.py # Semantic index orchestration and query
├── core/
│ ├── plugin_base.py # FeaturePlugin ABC — activate(), deactivate(), event helpers
│ ├── plugin_manager.py # Auto-discovery, loading, event bus, dock registry
│ └── events.py # Named constants for all plugin bus events
├── editor/
│ ├── ghost_editor.py # Editor with ghost text, minimap, inline chat, LSP
│ ├── multi_cursor.py # Multi-cursor editing logic
│ └── highlighter.py # Syntax highlighter registry
├── plugins/
│ ├── languages/ # Per-language syntax highlighting plugins
│ │ ├── python_plugin.py
│ │ ├── javascript_plugin.py
│ │ ├── typescript_plugin.py
│ │ ├── bash_plugin.py
│ │ ├── html_plugin.py
│ │ ├── markdown_plugin.py
│ │ ├── nix_plugin.py
│ │ ├── ansible_plugin.py
│ │ └── perl_plugin.py
│ ├── features/ # Auto-discovered feature plugins
│ │ ├── terminal/ # Embedded terminal dock (Ctrl+`)
│ │ │ ├── main.py # TerminalPlugin
│ │ │ └── terminal.py # TerminalDock, FallbackTerminal, QtermWidget
│ │ ├── import_graph/ # Import dependency graph visualization
│ │ │ ├── main.py # ImportGraphPlugin
│ │ │ └── import_graph.py # GraphDockWidget, GraphCanvas, force simulation
│ │ ├── symbol_outline/ # LSP-powered symbol outline panel
│ │ │ ├── main.py # SymbolOutlinePlugin
│ │ │ └── symbol_outline.py # SymbolOutlineDock
│ │ └── markdown_preview/ # Live markdown preview with scroll sync
│ │ ├── main.py # MarkdownPreviewPlugin
│ │ └── markdown_preview.py # MarkdownPreviewDock
│ └── themes/ # Theme definitions
│ ├── gruvbox_dark.py
│ ├── vscode_dark.py
│ ├── monokai.py
│ ├── solarized_dark.py
│ ├── solarized_light.py
│ ├── dracula.py
│ ├── nord.py
│ ├── one_dark.py
│ ├── palenight.py
│ └── quillai.py
└── ui/
├── menu.py # Application menus and recent projects
├── chat_renderer.py # Chat rendering, streaming, syntax highlighting
├── command_palette.py # Ctrl+P command palette
├── lsp_editor.py # LspEditorMixin — hover, go-to-def, squiggles, completions
├── breadcrumb_bar.py # File › class › method breadcrumb navigation
├── completion_popup.py # LSP completion dropdown with docstring preview
├── split_container.py # Split editor pane container
├── sliding_chat_panel.py # Sliding panel with Chat and Memory tabs
├── memory_manager.py # Per-project memory, facts, conversations, turns
├── memory_panel.py # Memory panel UI
├── git_panel.py # Source control panel
├── autosave_manager.py # Crash recovery and periodic autosave
├── startup_progress.py # Animated startup indicator
├── session_manager.py # Per-project tab session save/restore
├── find_replace.py # Find/replace panel
├── find_in_files.py # Project-wide search
├── snippet_palette.py # Snippet palette
├── settings_manager.py # Settings persistence
├── settings_dialog.py # Settings UI
├── diff_apply_dialog.py # AI rewrite diff preview
└── theme.py # Theme engine — stylesheet builders for all widgets
Create a folder under plugins/features/ with a main.py containing a FeaturePlugin subclass:
from core.plugin_base import FeaturePlugin
from core.events import EVT_FILE_OPENED
from PyQt6.QtCore import Qt
class MyPlugin(FeaturePlugin):
name = "my_plugin"
enabled = True
def activate(self):
from plugins.features.my_plugin.my_widget import MyDockWidget
self.dock = MyDockWidget(self.app)
self.app.my_dock = self.dock
self.app.addDockWidget(Qt.DockWidgetArea.RightDockWidgetArea, self.dock)
self.app.plugin_manager.register_dock("My Panel", "my_dock")
self.on(EVT_FILE_OPENED, self._on_file_opened)
def _on_file_opened(self, path=None, editor=None, **kwargs):
pass
def deactivate(self):
self.dock.close()
self.app.my_dock = NoneRestart QuillAI — the plugin is discovered and loaded automatically. See EVENTS.md for the full event reference.
All user data is stored locally:
| Data | Location |
|---|---|
| Settings | ~/.config/quillai/settings.json |
| Memory & facts | ~/.config/quillai/memory/ |
| Chat history | ~/.config/quillai/memory/chat_*.html |
| Sessions | ~/.config/quillai/sessions/ |
| Snippets | ~/.config/quillai/snippets.json |
| Vector index | ~/.config/quillai/vector/ |
| Autosave | ~/.config/quillai/autosave/ |
When using a local backend, no data is transmitted anywhere. When using a cloud backend, only the content you explicitly send is transmitted to that provider — nothing else.
- Code folding
- AI completion popup (Ctrl+Space for non-LSP files)
- Drag-and-drop tabs between split panes
- Git diff context in chat — auto-inject recent changes for debug queries
- Terminal stderr capture — pipe last error into chat context automatically
- Completion feedback loop — use acceptance data to influence suggestion ranking
- Plugin settings UI — enable/disable plugins at runtime from Settings dialog
- Plugin system — auto-discovery, event bus, dock registry; terminal, import graph, symbol outline, and markdown preview all implemented as plugins
- Split editor panes — horizontal and vertical, auto-collapse on last tab close
- Symbol outline panel — LSP-powered class/method tree with click-to-jump
- Import dependency graph — interactive force-directed visualization
- LSP completion dropdown — type signatures, docstrings, kind icons
- Breadcrumb bar — file › class › method navigation with symbol picker
- Markdown preview scroll sync — preview follows editor cursor and scroll position
- Smooth scrolling — ease-out wheel scroll animation
- Perl support — syntax highlighting, linting, LSP via perlnavigator
- LSP hover tooltips — formatted markdown with code block rendering
- Multi-cursor editing — Ctrl+D, Ctrl+Shift+L, Ctrl+Alt+Up/Down, Alt+Click
- Crash recovery — autosave every 2 minutes, silent restore on next launch
- Vector index — semantic search across code, conversations, completions, edit patterns
- LSP support — hover docs, Ctrl+Click go-to-definition, diagnostics, 8 languages
- Repo map — structural project index for codebase-aware chat
- Git blame in gutter
- Bracket match highlight
- Embedded terminal
- Command palette (Ctrl+P)
- Memory system with turn buffer and session continuity
- Line number double-click to select line
- LSP rename symbol
MIT