AI Potluck
Product / UX / Orchestration & agents

AutoGPT

Significant Gravitas

Autonomous agent platform where a user defines a goal and the agent breaks it into sub-tasks, executes them and adapts from the results. It began as a command-line experiment and is now centered on a low-code visual builder with a marketplace, where agents run on demand, on a schedule or from a trigger. Hosted and self-hosted deployment are both offered.

The repository ships two license files covering different parts of the tree, which the openness axis resolves. Verified 2026-08-13 via the Significant-Gravitas/AutoGPT repository.

Openness

2 high confidence
2.0
license
PolyForm-Shield(non-OSI, source-available, anti-compete
source
public

Dual-licensed: the current flagship, the AutoGPT Platform, is under PolyForm Shield - source-available, not OSI-approved, and anti-compete - while the legacy components (Classic, Forge, agbenchmark) stay MIT. The active product is what is scored. The README carries a per-component license table putting autogpt_platform/ under PolyForm Shield, described there as "free for personal and internal business use; cannot be sold as a competing hosted service", with the classic components under MIT; GitHub shows a bare "License" entry rather than a recognized SPDX badge, which is what a split license looks like to an automatic classifier. The whole platform source is published and self-hosting is offered with no license fee, so this reads as source-available: you can read it, but you cannot run it freely.

  • https://github.com/Significant-Gravitas/AutoGPT recorded 2026-08-13

    README license table - autogpt_platform/ under Polyform Shield, "free for personal and internal business use; cannot be sold as a competing hosted service", classic components under MIT; full source published with self-hosting available without a license fee; 186.6k stars, 46.1k forks

Adoption

3 low confidence
3.0

186,583 stars across the one declared repo, among the most-starred projects on GitHub, which bands at >10K stars, level 3 on the stars scale. That reflects enormous historical interest, but no platform download or monthly-active figure is published, and a star-based reading is capped at level 3 because a star is not a download. Recent production-usage signal is unclear, so the band is directional.

Capability

2 medium confidence
2.0

Pioneering but the autonomous loop has been surpassed; the low-code platform is its current value. No competitive agent-benchmark number published; scored low-mid on feature matrix.

Verified 2026-08-13