Azure DevOps Is Now a First-Class Dawn Plugin
Connect selected Azure DevOps projects, choose the agent actions you allow, and start Dawn workflows from pull request, work item, and pipeline events.
Read moreThe Dawn browser extension lets you give a Dawn agent access to browser pages. Bring the current tab into a thread, pin tabs for ongoing work, and drive them from any connected Dawn client even when the extension popup is closed.
Dawn Team
Product Engineering
Product updates, new features, engineering deep-dives, and guides for the Dawn AI agent platform.
Connect selected Azure DevOps projects, choose the agent actions you allow, and start Dawn workflows from pull request, work item, and pipeline events.
Read moreDawn Live Dashboards can now refresh governed server-side datasets from plugins, workspace files, personal connections, and public web pages.
Read moreDawn agents can retrieve and extract public web pages directly, while FourA remains available for use cases that require proxies or dedicated scraping infrastructure.
Read moreDawn now has a first-party plugin catalog for connecting workspace tools, choosing exactly what agents can do, and starting workflows from provider events.
Read moreDawn now supports Dropbox and Box as document integrations, so agents can browse, search, and read the files your team already keeps there, with folder-level control over what they see.
Read moreDawn now has /reflect, a slash command that audits recent workspace context and suggests memory updates, skills, commands, and workflows.
Read moreDawn now gives each user private memory pages for preferences, conventions, projects, decisions, and notes that should follow them across workspaces.
Read moreDawn can now create live dashboards from workspace data, files, connected sources, and agent-generated datasets so teams can inspect changing information without leaving the conversation.
Read moreDawn now supports xAI as a model provider, making Grok models selectable for agents alongside the rest of the Dawn model catalog.
Read moreDawn now includes an agent gallery so teams can create useful starter agents with recommended prompts, avatars, skills, and setup guidance instead of building every profile from scratch.
Read moreDawn now keeps large agent context more efficient by caching stable prompt sections and letting agents discover tools only when they need them.
Read moreDawn agents can use Google Analytics, Google Ads, and Google Search Console context to answer marketing, growth, and site-performance questions from workspace conversations.
Read moreDawn personal integrations let users connect Gmail, Google Calendar, Google Drive, and OneDrive so agents can help with private work context without making it workspace-wide.
Read moreAnthropic's most capable model is selectable for any Dawn agent from launch day, with improvements to long-horizon agentic coding, tool triggering, and reasoning effort calibration, plus a 1M token context window by default.
Read moreTwo new Google models are selectable for any Dawn agent: Gemini 3.5 Flash for fast multimodal work, and Gemini 3.1 Pro Preview for deeper reasoning on complex tasks. Both are available on every Dawn plan with no separate enable step.
Read moreDawn now makes files easier to mention, preview, inspect, and reuse across chat, generated outputs, screenshots, and agent-created artifacts.
Read moreDawn Workflows now give teams a visual canvas to design, validate, run, and inspect event-driven agent work, or have Dawn design the workflow itself from a plain-language description.
Read moreDawn agents can now run short JavaScript scripts inside a conversation, giving precise, deterministic answers for the data and file work that should not depend on the model getting it right.
Read moreDawn agents can now help teams create and update reusable Skills and Prompt Commands while keeping changes reviewable and protected from stale edits.
Read moreVoice Intelligence captures meetings, calls, and recorded conversations, transcribes them with ElevenLabs speech-to-text, separates speakers automatically, and turns the result into searchable, actionable workspace context.
Read moreDawn agents can now have custom avatars, including animated spritesheet packages you can generate with the new /avatar command - the next step in letting teams personalize how their agents show up.
Read moreDawn Workflows let teams run agents from connected-source events, starting with GitHub and Azure DevOps pull request review automation.
Read moreDawn now connects to FourA as a native integration. Assigned agents can scrape public web pages, extract structured content, and parse HTTP API responses — with the same per-agent governance as every other Dawn integration.
Read moreDawn now ships first-party integrations for both SendGrid and MailJet. Connect the workspace account, pick the scopes you want agents to use, and let the team find contacts, manage templates, and operate approved campaigns from the same conversations where the work is already happening.
Read moreEvery Dawn agent can now answer questions about Dawn itself — how workspaces, channels, integrations, commands, files, and tasks fit together — and turn the answer into the next step for your workspace. We are almost embarrassed it took this long.
Read moreDawn ships native integrations for the tools most teams already use. The new MCP Gateway is the answer for everything else — connect any hosted MCP server to your workspace, choose the capabilities, assign them to specific agents.
Read moreDawn Web now has a mini-games launcher in the status bar (and a /take-a-break command in chat). Eight retro arcade games for the moments when you just want a break while the agent gets on with it.
Read moreWorkspace owners can now allow Dawn to use provider-hosted web search, and the new /web-research command turns current public information into a focused result that fits the work in front of you.
Read moreDawn's global command pack now ships with focused prompt commands for PR reviews, ticket planning, release notes, and project briefs — ready to use, ready to copy into a workspace, and grounded in the connected sources teams already use.
Read moreDawn Web now has a dedicated files browser for generated outputs and thread attachments, so artifacts from past conversations are easy to find, preview, download, and reuse.
Read moreDawn now keeps long-running work visible in the thread as a background task, so you can leave it running, switch context, cancel it if it stops being useful, and apply the result when it is ready.
Read moreAccount-level personalization in Dawn now lets each user set a preferred name, a short bit of context about themselves, and a tone of voice — applied automatically whenever Dawn replies to a signed-in user.
Read moreA new /bug slash command in the Dawn CLI files a feedback report from inside your active session, attaching the current CLI logs and your workspace, agent, and thread context automatically.
Read moreThe new /explore command tells Dawn to investigate the workspace and connected sources in depth before answering, trading a few extra seconds for substantially better answers on broad or ambiguous questions.
Read moreGPT-5.5 is now selectable for any agent in Dawn, with reasoning support enabled. It uses three times the credits of GPT-5.4 per request, so it is best aimed at the work where the extra reasoning earns its keep.
Read moreDawn chat threads now run in parallel. Stop a run that's heading the wrong direction, move to another thread while a long one keeps working, and see at a glance which threads have a fresh reply waiting.
Read moreDawn CLI is now installable as a global Node command, giving teams an interactive terminal UI, non-interactive sends for scripts, local project context, attachments, and controlled shell access from the same Dawn agents they already use elsewhere.
Read morePrompt Commands let teams turn repeated Dawn workflows into reusable slash commands that carry the right instructions, use the current workspace context, and accept a short brief at runtime.
Read moreDawn agents can now read from Figma files and Framer projects, so design-aware answers and prototypes stay grounded in the visual system your team already ships.
Read moreOpenAI's freshly-released gpt-image-2 model is now wired into Dawn — ask an agent for an illustration, a hero image, a styled icon, or an edit of an existing picture and it produces one in the thread, alongside the SVG diagrams Dawn already drew.
Read moreAsk Dawn for a landing page, dashboard, status board, or lightweight prototype and it now returns a previewable HTML artifact you can open, revise, and download from the same thread.
Read moreScheduled tasks let Dawn run an agent prompt later or on a recurring schedule, then post the result back into the conversation or channel where the work belongs.
Read moreMemory pages give Dawn agents durable workspace context, so teams can keep decisions, runbooks, facts, and reference material in one place instead of repeating them in every thread.
Read moreDawn is now available as a first-party VS Code extension, so you can use Dawn in the editor, keep private threads separate from shared workspace chat, and let Dawn work against the codebase you already have open.
Read moreWorkspace owners can now link an existing OpenAI ChatGPT account so Dawn exposes those OpenAI models without consuming Dawn credits.
Read moreDawn workspaces now let owners add their own OpenAI, Anthropic, and OpenRouter API keys so teams can use their existing provider accounts without consuming Dawn credits.
Read moreDawn agents can now read from Confluence spaces and Notion page trees, so answers are grounded in the documentation your team already maintains.
Read moreDawn can now connect your Slack, Teams, Telegram, or Discord identity to your Dawn account, so the product recognizes you consistently no matter where you use it.
Read moreType / in Dawn to discover available commands. This release ships /help for guidance, /whats_new for changelog updates, and /link for identity linking — all discoverable directly from the chat composer.
Read moreDawn skills let teams capture how an agent should handle specific work — formatting, tone, checklists, document structure — and reuse it without restating the same prompt every time.
Read moreJira-connected answers fail less silently, Telegram conversations are more reliable, and integration setup tells you what went wrong instead of leaving you guessing.
Read moreDawn now works as an MCP server, so you can ask your Dawn agents questions directly from Codex, Cursor, and GitHub Copilot without leaving your editor.
Read moreUpload a PDF, spreadsheet, slide deck, or archive — Dawn can read it, extract what matters, and generate new files you can download and share.
Read moreDawn now supports private conversations across all channel types and a new teammate role that lets the wider team use Dawn without needing access to workspace administration.
Read moreDawn brings grounded answers about your own systems into Slack, Teams, Telegram, Discord, the web, and webhook-driven workflows.
Read more