aceteam-nodes 0.4.4

AceTeam workflow nodes for local execution

aceteam-nodes

AceTeam workflow nodes for local CLI execution. Run AceTeam workflows on your own machine with any LLM provider.

Install

pip install aceteam-nodes

Quick Start

# Run a workflow
python -m aceteam_nodes.cli run examples/hello-llm.json --input '{"prompt":"Explain AI in one sentence"}'

# List available nodes
python -m aceteam_nodes.cli list-nodes

# Validate a workflow file
python -m aceteam_nodes.cli validate examples/hello-llm.json

Configuration

Create ~/.ace/config.yaml to configure API keys and defaults:

default_model: gpt-4o-mini

Set your API key via environment variable:

export OPENAI_API_KEY=sk-...
# or
export ANTHROPIC_API_KEY=sk-ant-...

How It Works

Workflow JSON ──> AceTeamWorkflow (load + validate)
                       │
                       ▼
              workflow-engine (DAG execution)
                       │
                       ▼
              AceTeamNode.run() per node
                  │         │
                  ▼         ▼
              litellm    httpx/jinja2
            (100+ LLMs)  (API calls)

1. A workflow JSON file defines nodes and their connections as a directed acyclic graph
2. AceTeamWorkflow validates the file and resolves node references
3. The aceteam-workflow-engine executes nodes in topological order
4. Each node type implements run() — LLM nodes call litellm, API nodes use httpx, etc.
5. Results flow through the graph until all outputs are produced

Available Nodes

Node	Description
LLM	AI text generation via litellm (100+ providers)
APICall	HTTP requests with Jinja templating
Equal, NotEqual, GreaterThan, LessThan, And, Or, Not	Comparison/logic operators

Development

# Setup
uv sync --extra dev

# Test
uv run pytest

# Lint & format
uv run ruff check && uv run ruff format

# Type check
uv run pyright

# Build
uv build

Related

• Ace CLI — TypeScript CLI that wraps this package for a streamlined user experience
• Workflow Engine — The DAG execution engine used under the hood

License

MIT