mirror of https://github.com/getpaseo/paseo.git synced 2026-10-26 11:12:53 +00:00

Files

T

Mohamed Boudra 6e57bb1e4f fix(app): tell the truth about search scope and file identity

Two reproduced failures from the full correctness review.

An empty result claimed the whole workspace had been read. Ripgrep never
opens a file above the host's per-file ceiling, so a workspace whose only
match lives in a large log answered "No matches" with nothing else said, and
adding a small match still reported a complete result. The response now
carries the ceiling the host actually applied, and the empty state names it.
Only the empty state, because that is the only place the UI claims
completeness; a populated list already claims nothing, and the occurrence-cap
notice stays what it was. No permanent scope strip returns.

A file literally named "a\b.txt" previewed itself and then opened "a/b.txt".
The client rewrote every backslash to a separator on paths the host had
already made canonical, so a valid Unix file name became a directory and a
different file opened under a false "File changed" notice. Identity is now
kept as the producer gave it: the daemon renders workspace-relative paths with
its own separator turned into "/" for content search the way the file explorer
already did, and the client only normalises separators for a path whose own
shape is Windows, never from the client's operating system.

Native source previews sized every row at a hard-coded 20px while the row is
really one line at the reader's configurable code size, so scrolling to a
match landed in the wrong part of the file. The height is derived from the
size the view is already handed. Revealing a match thousands of lines into a
long file remains broken on Android beyond this; that is recorded with device
evidence rather than claimed fixed.

2026-09-10 11:45:15 +02:00

scripts

Fix graceful daemon shutdown on Windows (#4168 )

2026-09-01 19:59:03 +02:00

src

fix(app): tell the truth about search scope and file identity

2026-09-10 11:45:15 +02:00

.env.example

Use separate OpenAI endpoints for speech-to-text and text-to-speech (#1823 )

2026-06-30 10:00:27 +02:00

.gitignore

server(codex): normalize file paths, truncate large diffs, and drop turn/diff/updated telemetry

2026-02-06 11:51:59 +07:00

AGENTS.md

refactor(agent): add wait_for_agent tool and simplify prompt sending API

2025-11-03 11:47:15 +01:00

CLAUDE.md

Extract client SDK package (#1052 )

2026-05-28 01:58:18 +08:00

package.json

chore(release): cut 0.8.0-beta.1

2026-09-08 10:19:53 +02:00

README.md

chore: reformat codebase with oxfmt

2026-04-23 21:44:48 +07:00

tsconfig.scripts.json

Align script TypeScript lib with server build

2026-05-05 18:54:29 +07:00

tsconfig.server.json

2026-08-02 00:56:05 +08:00

tsconfig.server.typecheck.json

Fix repo typecheck failures

2025-11-15 20:17:47 +01:00

vitest.config.ts

Launch a terminal instead of a chat from New workspace (#2941 )

2026-08-07 14:08:22 +08:00

README.md

Voice Assistant

A voice-controlled terminal assistant that runs as a single local service.

Quick Start

# Install dependencies
npm install

# Copy environment variables
cp .env.example .env
# Edit .env and add your API keys (OpenAI, Deepgram)

# Run development servers
npm run dev

# Open browser to http://localhost:5173

Architecture

Express Server (port 3000) - Serves API and built UI in production
Vite Dev Server (port 5173) - Hot-reload React UI in development
WebSocket (/ws) - Real-time bidirectional communication
Agent - STT → LLM → TTS pipeline with terminal control
Daemon - tmux-based terminal management (in-process)

Development

# Run both servers (recommended)
npm run dev

# Or run separately:
npm run dev:server  # Express on port 3000
npm run dev:ui      # Vite on port 5173

# Type checking
npm run typecheck

# Build for production
npm run build

# Start production server
npm start

Project Status

✅ Completed (Phases 1-2):

Package setup and configuration
Express server with WebSocket
React UI with Vite
WebSocket client with ping/pong testing

⏳ In Progress (Phase 3):

Terminal control (tmux integration)

📋 Planned (Phases 4-9):

LLM integration (OpenAI GPT-4)
Agent orchestrator
Speech-to-Text (Deepgram)
Text-to-Speech (OpenAI)
Audio streaming
UI polish

See IMPLEMENTATION_PLAN.md for complete details.

Environment Variables

OPENAI_API_KEY=your-openai-key-here      # GPT-4 and TTS
DEEPGRAM_API_KEY=your-deepgram-key-here  # Streaming STT
STT_MODEL=whisper-1        # Optional: override to gpt-4o-transcribe, etc.
STT_CONFIDENCE_THRESHOLD=-3.0  # Optional: reject low-confidence clips
STT_DEBUG_AUDIO_DIR=.stt-debug # Optional: persist raw dictation audio for debugging
PASEO_HOME=~/.paseo        # Runtime state directory (agents/, etc.)
PASEO_LISTEN=127.0.0.1:6767  # Listen address (host:port or /path/to/socket)

PASEO_HOME defaults to ~/.paseo and isolates runtime artifacts like agents/. PASEO_LISTEN controls the daemon listen address. For blue/green testing you can run a parallel server without touching production state:

PASEO_HOME=~/.paseo-blue PASEO_LISTEN=127.0.0.1:7777 npm run dev

Tech Stack

Server: Express, TypeScript, ws (WebSocket)
Client: React 18, Vite, TypeScript
Terminal: tmux (via child_process)
AI: OpenAI (LLM + TTS), Deepgram (STT)

Testing

Currently manual testing via:

Start servers: npm run dev
Open http://localhost:5173
Test WebSocket connection (green status indicator)
Click "Send Ping" button to test communication

More testing guidance as features are implemented.

License

MIT