Keyword: AI
major
Agent runs: bound the pool of parallel tool calls (AgentLoop.TOOL_WORKERS) and tell a user whose chat message waits for a free worker
Agent screenshots belong to their chat: deleted with it, shown as thumbnails in the chat
tl-chat: Chat actions run as React commands within the window rather than via a separate ChatServlet; the chat's proxy workaround is no longer used in favor of engine polling
Ship a provider with recorded model answers for the agent tests of applications
TL Chat: Let a chat agent pass the input of a page command it runs
Adding Albert to an application: shipped React chat page, how-to article and archetype defaults
tl-ai-core pulls the Playwright browsers (driver-bundle, 200 MB) into every application: browser tools move into the optional module tl-ai-browser
AI Agents: Meaningful commit messages for changes made by an agent instead of "Agent edit"
major
TL Chat: While an agent is performing multiple steps, their last message remains visible, and the status bar displays the current step instead of just showing "…" and "Working on …"
TL Chat: A run that fails with an exception leaves no error in the conversation; it is gone after a reload
TL Chat: runPageCommand gives the model the refusal of a page command in English instead of the user's language
tl-ai-core, Anthropic provider: Claude 5.5 models missing from the bundled model catalog, and claude-haiku-5-5 sent a thinking budget (HTTP 400)
tl-ai-mcp-server: tools/list is not filtered by the caller's tool access, and /mcp/ with a trailing slash answers 404
tl-chat: A missing API key is reported as a rejected key ("The model rejected the access key")