RemoClaw v2.2.0 — Rebrand + landing page polish

🦞 The Chati → RemoClaw Release

This release is all about identity and presentation. New name, sharper positioning, and a lean README that actually reads like a developer landing page.

Why the rename?

"Chati" positioned this as a chat tool. The real value prop is remote control for AI coding agents — and IDEs increasingly want this but don't ship it. The new name + positioning makes that clear.

What's new

🎨 Rebrand

• All references updated: chati → remoclaw across code, scripts, configs, docs
• New tagline: Remote control any AI coding agent from your phone
• GitHub repo renamed: quangtam/chati → quangtam/RemoClaw
• Setup wizard banner shows RemoClaw FIGlet logo

📖 README overhaul (350 → 124 lines)

• Inspired by patterns from OpenCode, OpenHands, OpenClaw, Warp
• Hero → How it works → Quick start → Features → Docs flow
• Removed tagline duplication, v2.0 version labels (changelog noise)
• Inline navigation: Usage · Voice · Architecture · Deployment · Contributing
• Use cases section ("Code from your couch, fix bugs from the bus")

🌐 Vietnamese README

• New README.vi.md with full translation
• Language switcher at top of both English and Vietnamese versions

🎨 New SVG assets

• logo.svg — claw + signal waves brand mark
• how-it-works.svg — system diagram (phone → RemoClaw → AI agents)
• features.svg — 9-card feature map highlighting multi-project
• social-card.svg / social-card.png — 1280×640 Open Graph card for HN/Twitter/X shares

📚 Doc reorganization

• New docs/usage.md — full command reference + chat patterns (151 lines)
• New docs/voice.md — voice configuration + dual-backend details (96 lines)
• README points to docs/ instead of duplicating

🛠 Polish

• Updated CLAUDE.md agent context
• Updated all module docstrings
• Updated test file headers
• BMAD configs updated (project_name: RemoClaw)

Compatibility

• ✅ .env config unchanged — drop-in upgrade from v2.1.0
• ✅ SQLite schema unchanged
• ✅ All 342 tests pass
• ⚠️ Local file rename: chati.db → remoclaw.db on first start (fresh DB created)

Upgrading from v2.1.0

git pull
./remoclaw restart

That's it. Voice, multi-project, decision forwarding, screenshots — everything works the same.

---

Full Changelog: v2.1.0...v2.2.0

v2.1.0

What's New

• Auto-register bot commands — Bot commands now appear in Telegram's command menu automatically on startup
• Increased init timeout — CLI init timeout raised from 60s to 120s for slower connections

Chati v2.0.0 — Code from your pocket

What's New

6 epics, 24 stories, 342 tests.

🗄️ SQLite Persistence & Multi-Project (Epic 1)

• Thread configs, project bindings, and preferences survive bot restarts
• Bind different threads to different projects (/project, /projects)
• Per-thread provider switching (/provider claude)
• 3-layer config resolution: per-thread SQLite → global .env → defaults

⚡ Session Management (Epic 2)

• Concurrent session pool — multiple threads run in parallel
• Adaptive per-thread timeout
• Idle session auto-cleanup
• Streaming with progressive message edits (ChatGPT-like)

🔀 Interactive Decision Forwarding (Epic 3)

• Detects when CLI is waiting for input (Y/n prompts, file selections)
• Forwards decision prompts to Telegram with context
• Pipes your reply back to the CLI process
• Auto-kills sessions on reply timeout

📊 CLI Info & Session Visibility (Epic 4)

• /info — full session details (project, model, status, duration)
• /sessions — list all active sessions across threads
• Enhanced /status and /help

📸 Screenshot Forwarding (Epic 5)

• Auto-detects image paths in CLI output
• Sends as Telegram photos (or documents if >10MB)

🎤 Voice Communication (Epic 6)

• Voice input via Whisper (OpenAI cloud or faster-whisper local)
• Voice output via TTS (OpenAI TTS or edge-tts free)
• Auto-backend selection based on API key presence
• Code-heavy detection skips TTS for code responses
• Per-thread config: /voice, /voice speed, /voice status

🔧 Other

• /git command (branch, log, status, diff)
• BMAD skill routing
• Pluggable CLI drivers: Kiro, Claude Code, Gemini, Codex

---

Full Changelog: v1.0.0...v2.0.0

v1.0.1 — faster, parallel, resilient

What's new

Persistent interactive sessions (PTY)

CLI process stays alive across messages — no more cold start on every request.

• First message: ~6s (session init)
• Subsequent messages: ~3-4s (instant)
• Uses pseudo-terminal so CLI outputs in real-time

Per-thread parallel execution

Different Telegram threads now run concurrent CLI processes. No more waiting for another thread's request to finish.

Stuck detection & recovery

• Idle watchdog warns every 30s if no output (but doesn't kill — CLIs run subagents that buffer output)
• Global timeout (600s default, configurable via CLI_TIMEOUT)
• Partial output recovery if process dies mid-stream
• Typing indicator runs as background task — stays alive even when CLI is silent

Other improvements

• /cancel now works per-thread
• /status shows active session count
• Default timeout increased to 600s for complex tasks
• Typing keepalive independent of stream loop

Upgrade

git pull
./chati restart

v1.0.0 — code from your pocket

Chati v1.0.0

Chat with any AI coding CLI from your phone. No laptop needed.

Features

• Multi-CLI support — Kiro, Claude Code, Gemini CLI, OpenAI Codex
• Streaming output — real-time response with progressive message edits
• Thread = Session — each chat thread maps to a separate conversation
• Model selection — switch AI models via inline keyboard
• BMAD workflow — slash command routing for BMAD skills
• Markdown → HTML — CLI output converted to Telegram-native formatting
• Pluggable architecture — add a new CLI by writing 1 Python file
• Cross-platform — macOS, Linux, Windows

Quick Start

git clone https://github.com/quangtam/chati.git
cd chati
bash setup.sh    # Windows: setup.bat
./chati start    # Windows: chati start

Supported CLIs

Provider	Binary	Auth
Kiro	kiro-cli	kiro-cli login
Claude Code	claude	claude login
Gemini	gemini	gemini auth
Codex	codex	codex login

All CLIs use local login session. No API keys needed.

License

MIT — free to use, modify, and distribute.