10 mins read
OpenClaw v2026.3.24 landed on March 25, 2026, bringing a substantial round of platform upgrades, new integrations, and reliability fixes. This release focuses on enterprise messaging, AI gateway compatibility, and the skills management experience.
v2026.3.24 (2026-03-25) -- Includes Microsoft Teams SDK migration, OpenAI-compatible gateway endpoints, one-click skill installs, CLI container support, and a wide set of cross-platform stability fixes.
This release introduces several long-awaited capabilities, from AI-compatible API endpoints to a fully overhauled Microsoft Teams integration.
OpenAI-compatible endpoints: New
/v1/modelsand/v1/embeddingsgateway endpoints let RAG pipelines and tools built on the OpenAI SDK connect directly to OpenClaw.Microsoft Teams overhaul: Migrated to the official Teams SDK with support for streaming replies, welcome cards, AI response labeling, and message edit/delete actions.
One-click skill installs: Seven bundled skills now ship with install recipes. The Skills UI gains status tabs -- All, Ready, Needs Setup, and Disabled -- for at-a-glance management.
Container CLI support: The new
--containerflag andOPENCLAW_CONTAINERenvironment variable let you run OpenClaw commands inside Docker or Podman containers.
Run OpenClaw on Clawnify
Deploy your OpenClaw instance in minutes -- fully managed, always up to date.
A focused round of fixes across Discord, WhatsApp, Telegram, and the Node.js runtime keeps cross-platform behavior consistent and reliable.
Discord: Thread names are now auto-generated by the LLM. The
/toolscommand now reflects only tools available in real time.WhatsApp: Group echo suppression is restored and implicit reply detection is recovered after a regression.
Telegram: Fixed message routing for forum threads using the
#Generaltopic; added photo dimension preflight to prevent upload failures.Node.js: Minimum supported version is now 22.14+, with a preflight compatibility check that runs automatically before updates.
Security: Closed a sandbox media bypass that allowed
mediaUrlandfileUrlparameters to escape sandbox restrictions.
The direction of OpenClaw points toward deeper enterprise integrations, smarter AI routing, and a rapidly growing ecosystem of installable skills.
RAG and embedding pipelines: The new
/v1/embeddingsendpoint lays the groundwork for native retrieval-augmented generation workflows built directly on OpenClaw.Container-native deployments: The CLI container flag signals continued investment in Docker and Podman-first patterns for self-hosted and managed environments.
Skills ecosystem growth: With install recipes and status management now in place, the platform is primed for a broader catalog of first-party and community-contributed skills.
Smarter model routing: The v2 model router with keyword and NLI classification points to ongoing work on intent-aware, context-sensitive model selection across providers.
