AICTX is the official Python aictx CLI for operational continuity between AI coding-agent sessions.

Is AICTX an AI coding agent?

No. AICTX is not an autonomous coding agent. It is a repo-local continuity layer used by tools such as Codex, Claude Code, GitHub Copilot and generic coding agents.

What does AICTX preserve?

AICTX can preserve Work State, decisions, handoffs, known failures, execution summaries, contract signals, repo context and next actions across sessions.

Where is the official AICTX project?

The official AICTX website is https://aictx.org/, the official repository is https://github.com/oldskultxo/aictx, and the official Python package is aictx on PyPI.

Repo-local execution state for coding agents

Operational continuity for AI coding agents.

AICTX helps Codex, Claude, GitHub Copilot and other coding agents continue work across sessions by preserving the last useful execution state: active work, next actions, decisions, failures, validation evidence, repo context and quality signals for stale or unverified continuity.

Start in 2 minutes Open user manual View on GitHub

resume → work → finalize

pip install aictx
aictx install
aictx init

# Then just keep using your coding agent.
# Agents handle the lifecycle themselves:
# resume context → work → finalize continuity

Start from what actually happened.

Work State

Active task, hypothesis, relevant files, risks, next action and verification state.

Decisions and handoffs

Project decisions and what the previous session left for the next one.

Failures, validation and quality

Known failed commands, resolved patterns, canonical checks, validation evidence and stale-context quality signals.

Disconnected sessions are the problem.

Coding agents can reason through code, but every new session often starts operationally cold: it has to rediscover the repo, infer the task state, avoid old mistakes and guess which validation matters.

Without AICTX

A new agent session starts cold.
It scans README, docs and Makefile.
It rediscovers decisions.
It may repeat failed commands.
It guesses what should happen next.

With AICTX

The agent loads a resume capsule.
It sees active work and next action.
It sees known failures and validation path.
It finalizes factual evidence for the next session.

Inspectable, local, not hidden.

AICTX stores continuity under .aictx/. It is not cloud memory, not a vector database and not locked inside one chat. You can inspect it, clean it up, and optionally make a safe subset portable through Git.

Continuity View

Generate a local Markdown and Mermaid view of current operational continuity.

aictx view --repo .

# writes:
.aictx/reports/continuity-view.md
.aictx/reports/continuity-map.mmd

Local MCP server

Expose continuity as local MCP tools/resources/prompts. Compatible clients can launch it over stdio; CLI fallback remains available.

aictx mcp-server --repo . --profile full

Continuity Quality

Score whether repo-local continuity is fresh, stale, missing, demoted, obsolete or unverified before agents rely on it.

aictx doctor --repo . --json

# also exposed through:
# aictx resume --json
# MCP continuity quality

Simple start. Deep runtime when needed.

Quickstart

The shortest path from install to first continuity check.

Read Quickstart →

User Manual

Practical “I want to…” commands: portability, reports, MCP, cleanup, messages and diagnostics.

Open User Manual →

Technical Docs

Full technical reference for Work State, contracts, RepoMap, MCP, plugins, portability and internals.

Browse Technical Docs →

Works around the agents you already use.

Codex

CLI/runtime JSON contract, MCP support and Codex plugin artifacts.

Codex continuity →

Claude Code

Project instructions, hooks, MCP support and Claude Code plugin artifacts.

Claude Code continuity →

GitHub Copilot

Best-effort repo instructions, prompt files and VS Code MCP config where supported.

Copilot continuity →

What AICTX is not

AICTX is not an autonomous coding agent, a cloud memory service, a dashboard, a vector database or a replacement for human review.

AICTX does not guarantee correctness, productivity gains, token savings or that every agent will follow the lifecycle perfectly.