Least-privilege capabilities
A checkpoint link can only write. A separate continuation link can only read.
One history across every agent
Work history
that travels with you.
Save concise checkpoints after meaningful work, then bring the right decisions and next steps into a future session only when you ask. RecaPod keeps that private timeline outside any single AI provider.
One durable history. Separate read and write capabilities.
Free during early access · No subscription checkout is currently enabled.
Cross-agent work historyActive memory
Use one private RecaPod across every agent and session. Agents save meaningful work as structured checkpoints, then recall relevant history on demand. The agent must assert that you asked, and your client's approval setting remains the enforcing control:
Connected securely with MCP + OAuth
Latest checkpoint
Shipped the onboarding flow
Codex · 8 minutes ago
Ready anywhere
One timeline. Every agent.
Recall or export context on demand
The durable memory layer
Your work compounds across every session.
Structured, searchable history
Organize checkpoints by date, project, source, agent, session, tags, and status.
Recall starts locked
Agents save when instructed. Recall tools require the agent to assert that you asked; keep client tool approvals enabled because RecaPod cannot observe the conversation.
A concrete handoff
Save the outcome, not the whole chat.
A checkpoint captures the request, result, decisions, artifacts, project, status, and next steps. Later, you choose whether an agent should recall a focused slice.
See the full workflowConnected clients
Use the remote MCP endpoint with OAuth in clients that support authenticated custom servers.
Prompt-compatible agents
Use a write-only save prompt. Generate a separate 24-hour read link only for deliberate retrieval.
Example checkpoint
Shipped onboarding with mobile-safe authentication
Project: RecaPod · Status: completed · Artifacts: 3 · Next: verify production
Who RecaPod is for
Work that outlives one chat window.
Builders
Carry decisions, fixes, and next steps between coding agents.
Operators
Keep durable project context across recurring workflows.
Multi-agent teams
Give different AI clients a consistent, owner-controlled history.
Ready across every session.