Skip to content

AgentLoom

Deterministic LLM workflow orchestration with native observability, resilience, and cost control.

  • DAG-based workflows

    Define workflows as directed acyclic graphs in YAML or Python. Steps, dependencies, and routing are declared upfront — the LLM generates text, not control flow.

  • Native observability

    OpenTelemetry traces and Prometheus metrics on every step. Grafana dashboards included. No external SaaS required.

  • Built-in resilience

    Circuit breakers, rate limiters, and automatic multi-provider fallback. If OpenAI is down, the gateway falls back to Anthropic or Ollama.

  • Cost control

    Per-workflow budget limits, token tracking, and cost estimation across all providers. A workflow with budget_usd: 0.50 cannot overspend.

Installation

pip install agentloom

pip install agentloom[all]

pip install agentloom[graph]

Quick start

1. Create a workflow — my_workflow.yaml:

name: simple-qa
config:
  provider: openai
  model: gpt-4o-mini

state:
  question: "What is Python in one sentence?"

steps:
  - id: answer
    type: llm_call
    prompt: "Answer this question concisely: {state.question}"
    output: answer

2. Run it:

export OPENAI_API_KEY=sk-...
agentloom run my_workflow.yaml

agentloom run my_workflow.yaml --provider ollama --model phi4

export ANTHROPIC_API_KEY=sk-ant-...
agentloom run my_workflow.yaml --provider anthropic --model claude-sonnet-4-20250514

export GOOGLE_API_KEY=...
agentloom run my_workflow.yaml --provider google --model gemini-2.5-flash

3. Validate and visualize:

agentloom validate my_workflow.yaml    # check for errors
agentloom visualize my_workflow.yaml   # render the DAG

What's next

Section Description
Architecture Execution engine, DAG scheduler, state management
Providers Supported providers, models, and multi-modal capabilities
Workflow YAML Full reference for step types, config, routing, and attachments
Python DSL Build workflows programmatically
Graph API Analyze, visualize, and export workflow DAGs
Observability Traces, metrics, and Grafana dashboards
Examples 27 example workflows from basic to production-grade
Deployment Docker, Kubernetes, Helm, Terraform, and ArgoCD
Changelog Version history and release notes