[Release] openclaw/openclaw v2026.7.1-beta.5: OpenClaw 2026.7.1-beta.5

OpenClaw v2026.7.1-beta.5 introduces major features including conversational AI-guided onboarding through Crestodian, a new ClawRouter provider for dynamic model discovery and quota management, redesigned Control UI with session-first navigation, offline/spoken capabilities for mobile platforms, and support for GPT-5.6 models. The release emphasizes improved user experience across web, iOS, Android, macOS, and Apple Watch with features like auto-generated session titles, external harness attachment via Claude Code, and Telegram workflow enhancements.

Key Points

• •Conversational onboarding via Crestodian agent-loop with AI-guided provider setup, model-judged approvals, masked credentials, and deterministic fallback across CLI, web, and macOS
• •ClawRouter bundled provider enables credential-scoped dynamic model discovery with OpenAI-compatible and native Anthropic/Gemini transports plus managed budget reporting
• •Control UI redesign prioritizes sessions with minimal searchable sidebar, compact context ring, reasoning-effort slider, and native macOS session browser with model/thinking pickers
• •Offline and spoken mobile chat: iOS/Android pre-cache sessions per-gateway, Apple Watch supports full voice turns, iOS can speak replies via Gateway TTS with on-device fallback
• •Auto-generated session titles through utility-model routing with Gateway-backed groups, unread state, rename, fork, archive, and delete controls across all platforms
• •GPT-5.6 model family support across API-key, ChatGPT/Codex OAuth, and app-server paths with correct context, pricing metadata, and max reasoning support
• •External harness attachment: `openclaw attach` launches Claude Code against existing Gateway sessions with scoped, revocable, TTL-bound MCP grants and automatic revoke-on-exit
• •Telegram Codex workflows support private login, active run steering with `/steer` and `/tell` commands, and recovery across rejected rich entities and flood waits
• •Meta provider adds bundled `muse-spark-1.1` Responses API with streaming, tool calls, encrypted reasoning replay, and standalone npm/ClawHub distribution
• •Gateway improvements include host status display (network, OS, runtime, uptime, CPU, memory, disk), crash-loop recovery with control-plane-safe mode, and systemd/launchd integration

Workflow Diagram

Concepts

Artifacts (5)

messages.statusReactions.enabled: true

tts.speak - operator-scoped RPC that returns configured-provider speech as inline whole-clip audio for remote clients

openclaw attach

/login
/steer
/tell

@openclaw/meta-provider