aacp-crewai
AACP coordination layer for CrewAI multi-agent workflows.
CrewAI maps more naturally to AACP than any other framework because agent roles map directly to AACP domains and tasks map to AACP task types.
Install
pip install aacp-crewai
Quick start
from aacp_crewai.crew import AACPCrew
crew = AACPCrew(model="gpt-4.1-mini")
result = crew.run_workflow("payroll", period="2026-03")
print(result.summary())
Measured results
Payroll workflow comparison. gpt-4.1-mini. June 2026.
WITHOUT AACP WITH AACP
Coordination LLM calls: 4 0
Coordination cost: $0.0005 $0.0000
Total cost: $0.0005 $0.0003
Total saving: 39%
Coordination deterministic: NO YES
Schema validated: NO YES
The natural fit
CrewAI concept AACP concept
────────────── ────────────
Agent role DOM (HR, FIN, IT, SALES, CS)
Agent goal Workflow objective
Task description AACP packet content
Crew kickoff Orchestrator run
Without AACP vs With AACP
Without aacp-crewai (standard CrewAI):
Orchestrator
↓ "Retrieve all active employee salary records for March 2026.
Include employee ID, name, department, cost centre, base salary,
any changes this month, and pension rate. Return as JSON."
HR Agent ← verbose, varies every run
With aacp-crewai:
Orchestrator
↓ FETCH|HR|return:ORCHESTRATOR|p:1|aacp:1.1|res:emp_salary|period:2026-03
HR Agent ✓ validates. $0.00 encoding. Identical every run.
Comparison demo
python3 examples/comparison.py --mock
Workflows
# Payroll (5 hops)
crew.run_workflow("payroll", period="2026-03")
# IT provisioning / JML (6 hops)
crew.run_workflow("it_provisioning", username="j.smith", dept="Engineering")
# Sales qualification (5 hops)
crew.run_workflow("sales_qualification", lead_id="L-001")
Requirements
- Python 3.10+
OPENAI_API_KEYpip install aacp-crewai
Links
- Protocol spec: https://aacp.dev
- Python SDK: https://github.com/MackayAndrew/aacp
- LangChain integration: https://github.com/MackayAndrew/aacp-langchain
- Community rules (241): https://github.com/MackayAndrew/aacp-community-rules
- IETF Draft: https://datatracker.ietf.org/doc/draft-mackay-aacp/
Licence
MIT
aacp-crewai
AACP coordination layer for CrewAI multi-agent workflows.
CrewAI maps more naturally to AACP than any other framework because agent roles map directly to AACP domains and tasks map to AACP task types.
Install
pip install aacp-crewai
Quick start
from aacp_crewai.crew import AACPCrew
crew = AACPCrew(model="gpt-4.1-mini")
result = crew.run_workflow("payroll", period="2026-03")
print(result.summary())
Measured results
Department day comparison. 5 workflows. 59 coordination hops. gpt-4.1-mini. June 2026.
WITHOUT AACP WITH AACP
Coordination LLM calls: 59 0
Coordination cost: $0.0008 $0.0000
Agent cost: $0.0018 $0.0018
Total cost: $0.0025 $0.0018
Total saving: 30%
Coordination deterministic: NO YES
Schema validated: NO YES
Audit trail structured: NO YES
CrewAI's natural language task descriptions are more verbose than LangChain's by default, which is why the per-hop saving is higher. Both comparisons use the same 59-hop department day scope for a fair comparison.
The natural fit
CrewAI concept AACP concept
────────────── ────────────
Agent role DOM (HR, FIN, IT, SALES, CS)
Agent goal Workflow objective
Task description AACP packet content
Crew kickoff Orchestrator run
Without AACP vs With AACP
Without aacp-crewai (standard CrewAI):
Orchestrator
↓ "Retrieve all active employee salary records for March 2026.
Include employee ID, name, department, cost centre, base salary,
any changes this month, and pension rate. Return as JSON."
HR Agent ← verbose, varies every run
With aacp-crewai:
Orchestrator
↓ FETCH|HR|return:ORCHESTRATOR|p:1|aacp:1.1|res:emp_salary|period:2026-03
HR Agent ✓ validates. $0.00 encoding. Identical every run.
Comparison demo
python3 examples/comparison.py --mock
Workflows
# Payroll (5 hops)
crew.run_workflow("payroll", period="2026-03")
# IT provisioning / JML (6 hops)
crew.run_workflow("it_provisioning", username="j.smith", dept="Engineering")
# Sales qualification (5 hops)
crew.run_workflow("sales_qualification", lead_id="L-001")
Requirements
- Python 3.10+
OPENAI_API_KEYpip install aacp-crewai
Links
- Protocol spec: https://aacp.dev
- Python SDK: https://github.com/MackayAndrew/aacp
- LangChain integration: https://github.com/MackayAndrew/aacp-langchain
- Community rules (241): https://github.com/MackayAndrew/aacp-community-rules
- IETF Draft: https://datatracker.ietf.org/doc/draft-mackay-aacp/
Licence
MIT
Metadata
Release files for aacp-crewai 0.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| aacp_crewai-0.1.0.tar.gz | 16.0 kB | Without provenance | Details for aacp_crewai-0.1.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| aacp_crewai-0.1.0-py3-none-any.whl | Python 3 | none | any | Details for aacp_crewai-0.1.0-py3-none-any.whl |
Total release size: 34.7 kB
Release files / aacp_crewai-0.1.0.tar.gz
| Download URL | aacp_crewai-0.1.0.tar.gz |
|---|---|
| Size | 16.0 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
3171e87889213976f74710172db09d5df9e7f97914f617878aba68639f9a82b8
|
|
BLAKE2b-256 checksum How to use checksums |
e987976c3badeb597eed81f64ffed5730b5e186caa6b60749eea9636d5d1c739
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.3
|
Release files / aacp_crewai-0.1.0-py3-none-any.whl
| Download URL | aacp_crewai-0.1.0-py3-none-any.whl |
|---|---|
| Size | 18.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
dd17de7ee80f24777fe7c697b9e5f90ad83fdf37ea6e3fb150eca3ae1fc0bab1
|
|
BLAKE2b-256 checksum How to use checksums |
e45311055352b026a437ed82f0ab3be290bbb3c82bdc9e15fb88a85124cde560
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.3
|
Security for aacp-crewai 0.1.0
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.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.