[Release] openclaw/openclaw v2026.7.1-beta.2: openclaw 2026.7.1-beta.2
OpenClaw v2026.7.1-beta.2 introduces GPT-5.6 support, external harness attachment via `openclaw attach`, Telegram Codex workflows with login and steering, event-driven cron runs with `on-exit` scheduling, and native app improvements including iOS 26 visual system adoption. The release also adds ClawRouter for dynamic model discovery with credential-scoped routing, richer iMessage messaging with native polls, safer scoped conversations via capability profiles, and enhanced Control UI navigation with reasoning-effort controls. Extensive fixes address Telegram durability, agent reliability, provider safety, channel delivery, cron correctness, and memory recovery across multiple platforms.
Key Points
- •GPT-5.6 model family now recognized across catalog, capability, and runtime selection paths for broader LLM support
- •`openclaw attach` command enables external harness attachment to existing Gateway sessions for resumable Codex-style workflows
- •Telegram integration supports `/login` for Codex pairing, active run steering, and transient API failure recovery
- •Event-driven cron with `on-exit` schedule kind triggers agent execution when watched commands exit; session-targeted runs detach cleanly
- •ClawRouter provider plugin delivers credential-scoped dynamic model discovery, OpenAI-compatible and native Anthropic/Gemini transports, and managed budget reporting
- •iOS adopts iOS 26 visual system with modernized Chat, Talk, onboarding, and reconnect flows; macOS app auto-installs local Gateway
- •iMessage gains native poll creation/voting and built-in usage footers; Telegram adds multi-lane progress summaries
- •Scoped conversation capability profiles enable per-conversation tool and access boundaries without weakening default profiles
- •Control UI redesign includes session-first sidebar, compact context meter, warm light theme, reasoning-effort slider, and slash-command picker
- •Comprehensive fixes for Telegram durability (stalled ingress, media retry, poison updates), agent reliability (runtime overrides, context estimation), and provider safety (response bounds across 10+ platforms)
Found this useful? Add it to a playbook for a step-by-step implementation guide.