# prefect-archive/ControlFlow

🦾 Take control of your AI agents

Repository: https://github.com/prefect-archive/ControlFlow
Canonical: https://ross.abutalabs.com/products/controlflow
Homepage: https://controlflow.ai
Language: Python
License: Apache-2.0
License Family: permissive
Topics: agentic-framework, agentic-workflow, ai, prefect, workflow, agent, llm
Archived: true
Last push: 2025-08-22T13:51:57+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 38, release rhythm 40, longevity 63
- inputs: {"age_days": 884, "days_push": 376, "days_rel": 573, "gap_med": 5, "n_releases_24m": 14}
- flags: archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1390, forks 111 (observed 2026-08-28T04:04:35.548913+00:00)

## What it is
ControlFlow is a Python framework for building agentic AI workflows, letting developers define discrete observable tasks, assign specialized LLM-powered agents, and orchestrate them in flows with type-safe structured outputs. It integrates with Prefect 3.0 for observability, but the project has been archived and its next-generation engine merged into the Marvin framework.

## Use cases
- build multi-step LLM agent workflows in python
- orchestrate multiple specialized ai agents on tasks
- get type-safe structured outputs from llm calls
- add observability to ai agent pipelines with prefect
- delegate work to llms with controlled autonomy
- prototype agentic ai applications quickly

## When to choose
- you want a task-centric python framework for llm agents with structured results
- you already use prefect and want native observability for ai workflows
- you need fine-grained control over agent autonomy

## When to avoid
- you need an actively maintained project - the repo is archived and development moved to Marvin
- you want a non-python or low-code agent builder
- you need provider support beyond what its llm configuration offers

## Facets
- artifact type: framework
- maturity: abandoned
- function: agent-framework, llm-inference, workflow-automation, prompt-engineering
- domain: large-language-models, developer-tools
- platform: python, cross-platform
- tags: agentic-framework, multi-agent-orchestration, task-centric, structured-outputs, prefect, archived, observability, ai-agents, automation

## Member repositories
- prefect-archive/ControlFlow (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:35.548913+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:39:39.983371+00:00, confidence not recorded.
  - readme: https://github.com/prefect-archive/ControlFlow (fetched 2026-08-28T04:04:35.548913+00:00, sha 0c395affa1d5)
- Data as of 2026-08-30T08:39:29.467469+00:00.