Recall turns your private work sessions into structured context for the AI tools you choose, connecting the project, file, tab, issue, message, open question, and next step behind your work.
Your tools hold the pieces. Recall understands which pieces belonged together.
Even when an agent connects to your stack, it usually starts from separate tools: one Slack thread, one GitHub PR, one Notion doc, one Figma frame, one browser tab. It still does not know what you were doing, why those pieces mattered, or where you stopped.
Recall is the context router between your work stack and the AI agents you choose.
Recall watches the shape of work locally, groups fragments into projects, threads, and tasks, and can turn that context into a private Context Card for the AI tools you approve.
Exploring. Tools shown are examples of the stack Recall reads locally, not a list of live connections.
Five jobs Recall context could power, with the AI tools you already use.
Ask Claude, Codex, Cursor, or your own agent for help without rebuilding the whole setup. Recall can hand over the active project, thread, relevant artifacts, open question, and next step.
Recall can group the Slack message, Linear issue, GitHub PR, Figma frame, Notion note, and browser tab that belonged to the same piece of work.
Instead of asking an agent to guess, Recall can point it to the exact file, tab, doc, issue, frame, PR, or message thread behind the task.
Generate standups, founder updates, project summaries, client notes, or end-of-day recaps from the sessions you actually moved, including decisions, artifacts, open loops, and next steps.
Recall's session state can help agents know when to stay quiet, when to prepare context, and when to offer one clear anchor: Locked In, Scattered, Re-entry, Overloaded, or Reactive.
Recall is exploring a local MCP server that lets approved AI clients request structured work context from Recall.
get_current_threadThe active project, thread, and artifacts.get_return_cardWhere you stopped and the next step.list_open_loopsUnfinished threads still waiting on you.summarize_dayWhat actually moved today.summarize_projectChanges across a project over time.search_threadsFind a past work thread by intent.get_focus_stateCurrent session state, for tone and timing.open_return_pointJump back to the exact artifact.Pre-built workflows that use your Recall context through MCP. Pick the outcome, connect your approved AI tool, and run the skill.
A Recall Skill is not a generic prompt. It combines a job to be done, the Context Card fields it needs, your connected stack tools, the expected output, and the AI clients it works with.
Turn your active Recall thread into a clean brief for Claude, Codex, Cursor, or your own agent.
Use the return card, open question, and next step from an unfinished thread.
Generate a daily update from the sessions you actually moved.
Summarize what changed inside a project over the last day, week, or sprint.
Load the file, issue, PR, failing test, and last open question behind a coding thread.
Give your agent the Figma frame, feedback thread, related issue, and notes behind a design review.
Show what you were working on before a meeting broke the flow, and prepare the cleanest next step.
Use today's open loops and warmest unfinished sessions to prepare the best first work block for tomorrow.
Want early access to Recall MCP and Skills?
Join the waitlistRecall's agent layer keeps the same promise as the core product: your work context is for you. It travels to an agent only when you choose to share it.
Read the privacy principles →Recall helps you return to your work today. Soon, it can also hand that thread to the AI tools you trust.