Skip to main content

AWR

Let your AI team keep complex projects moving—from goal to verified delivery.

The open-source project delivery platform for people and AI.

Website · Documentation · 简体中文 · Release notes

AWR connects goals, workstreams, dependencies, context and acceptance into one project delivery workflow. Keep the project's intent, constraints and verified progress when you change sessions or agents. People set direction and review outcomes; agents work through the same project using CLI or MCP.

AWR architecture: project files remain authoritative; the runtime connects project state, focused context, checkpoints and evidence to people and agents.

What 0.5.1 provides

This package installs the native Rust awr CLI and awr-mcp server for personal, local SQLite projects. No model call is needed to compile context.

  • Continue across sessions. Keep goals, rules, decisions, claims and checkpoints connected to the current work.
  • Use the context that matters. Compile task-specific project context within a budget while retaining required facts and exposing missing dependencies.
  • Keep progress honest. Distinguish source status, checkpoint observations and version-bound verification evidence; retain caller attribution.
  • Check source freshness consistently. Build a shared file inventory that ignores .DS_Store and other supported metadata, and records real file changes.
  • Organize personal workstreams. Explicit source ownership, session attribution and scoped context form the foundation for parallel work.
  • Use your existing agent. CLI, stdio MCP and a shared HTTP MCP service connect supported clients and projects; optional lifecycle adapters deepen integration.

The optional Inspector runs from source and is not bundled in npm/PyPI. This release's source includes Chinese/English views, complete queue pagination and session/event panels.

Team/PG collaboration, advanced delivery adoption, DEC extensions, usage and ETA accounting are outside 0.5.1. The personal workstream foundation is not a claim of complete authenticated Team isolation. See the workstream guide for the supported boundaries.

Install with pip

Requires Python 3.9+. Use a virtual environment:

python -m venv .venv
# Activate .venv using your shell's standard command, then:
python -m pip install agent-work-runtime==0.5.1
awr --version
awr-mcp --version

For an isolated application installation, use pipx install agent-work-runtime==0.5.1 or uv tool install agent-work-runtime==0.5.1.

Wheels embed both native binaries; Rust is not needed for installation. Python provides launchers rather than a separate Python SDK. Source identity, binary hashes and third-party license notices travel with the native payload.

Connect a project

From your project directory, preview initialization and review the proposed sources before accepting the same goal:

awr init --goal "Deliver a customer portal with verified sign-in"
awr init --goal "Deliver a customer portal with verified sign-in" --accept
awr status
awr intake inspect

Existing Markdown/YAML sources remain authoritative. If intake reports NeedsOrganization, fill in the required goal, work and acceptance information before implementation. See the project intake guide.

Give your agent this working agreement:

Use AWR to carry this project from its goal to verified delivery. Read the current goals, rules, dependencies and checkpoint before starting. Keep clear acceptance criteria and ownership. Record actual progress and evidence, and save a checkpoint before stopping so the next session can continue.

Connect through MCP

After initialization, add this entry using your client's supported MCP configuration format and your project's absolute path:

{
  "mcpServers": {
    "awr": {
      "command": "awr-mcp",
      "args": ["--project", "/absolute/path/to/your/project"]
    }
  }
}

Reconnect the client and verify the project identity. For multiple clients and projects, see the shared HTTP MCP service guide. CLI/MCP is the common entry point for Codex, Cursor, Claude Code and other compatible agents. Host hooks are optional and require their own activation; a saved configuration does not prove a checkpoint was executed.

Supported systems and upgrades

Prebuilt packages support macOS 15+ (Apple Silicon and Intel), Linux x64/arm64 with glibc 2.39+, and Windows x64. SQLite is bundled; Git-bound operations also need Git. Other platforms require a separately verified source build. See the distribution guide.

Upgrading from 0.5.0: stop writers and retain matching program, runtime and source snapshots first. Explicit source refresh upgrades SQLite schema 4 through 5 and 6 to 7, preserving legacy identities and history. Old binaries refuse schema 7. A rollback requires the matching pre-upgrade snapshot, not just a package downgrade; keep any post-upgrade history separately.

Learn more

Licensed under Apache-2.0.

Release files for agent-work-runtime 0.5.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distributions (wheels)

Table of built distributions (wheels) for agent-work-runtime 0.5.1
File
agent_work_runtime-0.5.1-py3-none-win_amd64.whl Python 3 none Windows x86-64 Details
agent_work_runtime-0.5.1-py3-none-manylinux_2_39_x86_64.whl Python 3 none Linux glibc 2.39+ x86-64 Details
agent_work_runtime-0.5.1-py3-none-manylinux_2_39_aarch64.whl Python 3 none Linux glibc 2.39+ ARM64 Details
agent_work_runtime-0.5.1-py3-none-macosx_15_0_x86_64.whl Python 3 none macOS 15.0+ x86-64 Details
agent_work_runtime-0.5.1-py3-none-macosx_15_0_arm64.whl Python 3 none macOS 15.0+ ARM64 Details

Total release size: 106.8 MB

Release files / agent_work_runtime-0.5.1-py3-none-win_amd64.whl

Download URL agent_work_runtime-0.5.1-py3-none-win_amd64.whl
Size 20.2 MB
Tags Python 3 Windows x86-64
SHA-256 checksum
How to use checksums
a3be0fe5f29b9925887f3efdd7dec15cc183bfd67575667de739f04b71f73618
BLAKE2b-256 checksum
How to use checksums
ba30fb32526e4a1aa56c5694bad934800ec4c00590d7084a85e3923b31abc291
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 26, 2026.

Transparency log

Release files / agent_work_runtime-0.5.1-py3-none-manylinux_2_39_x86_64.whl

Download URL agent_work_runtime-0.5.1-py3-none-manylinux_2_39_x86_64.whl
Size 22.5 MB
Tags Linux glibc 2.39+ x86-64 Python 3
SHA-256 checksum
How to use checksums
38134e5dab53c5f531ec94f3522316c7a3d3090f0a4ec9c3f13bdb924948d54f
BLAKE2b-256 checksum
How to use checksums
f49c804b532e8894314b8af79304a54327dec0756fc9d58f82ec5dcd5bb2c7c9
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 26, 2026.

Transparency log

Release files / agent_work_runtime-0.5.1-py3-none-manylinux_2_39_aarch64.whl

Download URL agent_work_runtime-0.5.1-py3-none-manylinux_2_39_aarch64.whl
Size 21.8 MB
Tags Linux glibc 2.39+ ARM64 Python 3
SHA-256 checksum
How to use checksums
6742b8e5d3acba075148bf5441b5fda5df63d866d70f6d39880158903ffa565b
BLAKE2b-256 checksum
How to use checksums
7bde8d8d8be06ef1c0ee52f9f9465cfb64e2d2f374616a9d37f7a8e5bb717c4b
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 26, 2026.

Transparency log

Release files / agent_work_runtime-0.5.1-py3-none-macosx_15_0_x86_64.whl

Download URL agent_work_runtime-0.5.1-py3-none-macosx_15_0_x86_64.whl
Size 21.6 MB
Tags Python 3 macOS 15.0+ x86-64
SHA-256 checksum
How to use checksums
38099456c125682a47779b4cf3cc702ad9a9b968bac221b979d0c1d1c54bb36f
BLAKE2b-256 checksum
How to use checksums
7c093fa2feec1fad5c913587e98ab65982d3d28700f9ccf0ef94cff54177a42b
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 26, 2026.

Transparency log

Release files / agent_work_runtime-0.5.1-py3-none-macosx_15_0_arm64.whl

Download URL agent_work_runtime-0.5.1-py3-none-macosx_15_0_arm64.whl
Size 20.6 MB
Tags Python 3 macOS 15.0+ ARM64
SHA-256 checksum
How to use checksums
8ad885332c936d11171ddcfbc4dce1bf713ca28541380a5244941ef47d01911b
BLAKE2b-256 checksum
How to use checksums
3cfc303051d5061552e9db5c4604055a70f765e8d4df8ac78b50a6b64a82615c
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 26, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

0.5.1 This release

5 release files

0.5.0

5 release files

0.4.0

4 release files

0.3.3

4 release files

0.3.2

4 release files

0.3.1

4 release files

0.3.0

3 release files

0.2.0

3 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page