tramAI vs shepherd
AI workflow governance and agent supervision compared
tramAI provides a Kotlin-first JVM runtime for typed AI workflows, with configurable policy, provider routing, approvals, and execution evidence. shepherd is a Python framework that records agent runs as reversible traces and keeps workspace changes as proposals for review. tramAI focuses on governing model use within an application, while shepherd focuses on supervising agents and their changes.
tramAI: Govern AI Workflows on the JVM
TramAI is a Kotlin-first JVM runtime for typed AI workflows with configurable policy, provider routing, tool controls, approvals, and execution evidence. It suits teams that need governance around model use, not just a way to call an LLM.
shepherd: Supervise Agents with Reversible Execution Traces
Shepherd records agent work as inspectable, reversible execution traces and keeps changes as proposals for review. It is aimed at developers building systems that supervise, replay, or manage the work of other agents.
| tramAI | shepherd | |
|---|---|---|
| Language | Kotlin | Python |
| License | Apache-2.0 | MIT |
| Stars | 35 | 2.5k |
| Forks | 0 | 217 |
| Last analyzed | Oct 9, 2026 | Oct 4, 2026 |
Key differences
- tramAI is written in Kotlin for JVM applications; shepherd is a Python framework for agent tasks in controlled workspaces.
- tramAI governs AI operations through configurable policy, data controls, provider routing, and approval capabilities; shepherd manages agent-run traces and proposed workspace changes.
- tramAI supports standalone and Spring Boot integration and lists several provider adapters; shepherd offers repository permissions and a deterministic provider for trying its retained-output workflow.
- tramAI uses the Apache-2.0 license; shepherd uses the MIT license.
- tramAI describes its 0.6.x line as frozen for bug and security fixes, with new features planned for 0.7.0; shepherd is in early alpha and says its APIs may change.
- tramAI is aimed at JVM teams building governed AI workflows; shepherd targets developers building systems to supervise, replay, or manage other agents.
Choose tramAI if you…
- need typed AI workflows within a Kotlin or Spring Boot application.
- want configurable policy, provider routing, or approval gates around model and tool use.
- need execution evidence for AI workflows, with persistence and recovery options when configured.
Choose shepherd if you…
- want to inspect agent runs and review proposed workspace changes before applying them.
- are building meta-agents or other systems to supervise, replay, or compare agent execution.
- want to explore trace and proposal workflows with a deterministic provider.
Compare the numbers side by side, or add more projects
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.