stn.
Get Started Free

Overview

stndp is shared, durable memory for engineering teams and their AI agents, driven from the terminal with the stn CLI. It captures the decisions, working state, and reasoning behind your work, and lets your coding agents read and write that same memory over MCP. stndp extends the memory your agents already have — each tool's built-in memory is per-session and per-machine, while stndp is the shared layer that persists across sessions, agents, and teammates — so context written once is there for the next person, and the next agent.

Install

pip install stndp-cli   # provides the `stn` command (Python 3.11+)

Then authenticate in your browser:

stn login

Start here

  • Quickstart — from install to your first captured context (and your first reload-able checkpoint) in a few minutes.
  • How stndp works — the memory types and the context graph behind them.
  • CLI reference — every command and option, split by topic (account, updates, checkpoints, decisions, episodes & bugs, the graph).
  • Coding agents (MCP) — wire stndp into Claude Code, Cursor, Codex, Gemini CLI, or Windsurf so they remember across sessions.
  • Integrations and webhooks — GitHub, Jira, Slack, and CI events.

What you can do

  • Resume where anyone left off — stn checkpoint saves your private working state (focus, why, next step, git context); stn resume reloads it — for you or your agent — at the start of the next session.
  • Record decisions and learnings — stn decide captures a durable, searchable rationale; stn search finds anything across your team's shared memory.
  • Capture the arc of work — stn episode records what you planned, what happened, and the surprise; stn bug files a symptom → cause → fix a future search can find.
  • Keep the team in sync — stn push TEAM "..." posts an update; --blocked raises a blocker that @mentions whoever can clear it.
  • Plug in your agents — stn agent setup wires the stndp MCP server into Claude Code, Cursor, Codex, Gemini CLI, or Windsurf so they resume, search, and log to the same memory you do; stn guide is the workflow playbook they (and you) can print any time.

The shape of your memory

stndp stores a handful of memory types, each with a clear home:

  • Entries — team-visible, dated updates and blockers.
  • Checkpoints — private working state, captured at context switches and reloaded with resume.
  • Decisions — durable, team-visible choices and learnings, searchable forever.
  • Episodes — the arc of a unit of work: plan → outcome → the surprise that becomes the lesson.
  • Bugs — first-class symptom → cause → fix records, so the next person (or agent) who hits one gets a straight answer.

Everything connects into one context graph — repos, files, tickets, PRs and people — that both you and your agents can query. Your context is yours: export it any time with stn export.

Help

Run stn guide for the workflow playbook, or stn --help / stn COMMAND --help for any command. Questions or issues? Email support@stndp.io, or see the Trust Center for security and data questions.