mngrvsXum
Generated from the two spec rows. Green marks the better value where a spec has a clear direction. Everything else is just different.
mngr
Imbue · Agent harness
OSS
Panel
7.12 spec wins
- Reliability
- 6.7
- Usefulness
- 7.3
- Cost
- 8.3
- Longevity
- 6.0
“It gives you push, pull, clone and snapshot for agents, so you can now have a merge conflict with a colleague you invented.”
Xum
Coder · Agent harness
OSS
Panel
6.82 spec wins
- Reliability
- 6.3
- Usefulness
- 6.3
- Cost
- 8.3
- Longevity
- 6.2
“It renders mermaid diagrams and LaTeX in your terminal, so the agent can now hand you an architecture you did not ask for, beautifully.”
Spec by spec
| Spec | mngr | Xum |
|---|---|---|
| Architecture | ||
| Category | Agent harness | Agent harness |
| Runs | local, cloud, sandbox | local |
| Platforms | macos, linux | macos, linux |
| Context window | not documented | not documented |
| Protocols | ||
| MCP client | No | No |
| MCP server | No | No |
| Capabilities | ||
| Runs terminal commands | Yes | Yes |
| Multi-file edits | Yes | Yes |
| Git operations | Yes | Yes |
| Browser control | No | No |
| Sandboxed execution | YesAgents can be placed in Docker containers, on Modal or on remote hosts, with SSH key isolation, network allowlists and full container control. | No |
| Multi-agent orchestration | Yes | Yes |
| Headless / CI mode | Yes | No |
| Models | ||
| Backbone | via managed agents (Claude Code, Codex, OpenCode) | Anthropic, OpenAI, Grok, Ollama, OpenRouter |
| Bring your own model | Yesmngr ships no model; the model comes from whichever agent CLI it launches. | Yes |
| Local models | No | Yes |
| Cost | ||
| Pricing model | byok | byok |
| Starts at | $0/mo | $0/mo |
| Free tier | Yes | Yes |
| Bring your own key | Yes | Yes |
| Openness | ||
| Open source | Yes | Yes |
| License | MIT | AGPL-3.0 |
| GitHub stars | 411 | 2,045 |
Which one would each critic pick
| Critic | mngr | Xum | Pick |
|---|---|---|---|
| El Juez | — | — | not enough reviews |
| El Amigo | 7.5 | 7.3 | mngr — Pick it when you are running more agents than you can remember; pick a single terminal session if you have never once wondered what an agent was waiting for. |
| El Crítico | 6.8 | 6.3 | mngr — It is built on four systems it does not own, and a failure is one of four failures, logged in four places, none of which is the tool you were using at the time. |
| El Profesor | 6.8 | 6.8 | no preference |
| La Inversora | 6.5 | 6.5 | no preference |
| La Jefa | 7.0 | 6.3 | mngr — Key isolation per agent, network allowlists and full container control across sixty engineers, and the cost is compute rather than seats, which is a line I already forecast. |
| El Hacker | 8.0 | 7.8 | mngr — MIT, plugins to extend it, and the same commands drive Claude Code, Codex or OpenCode, so the harness stops being a thing I choose once and then live with forever. |
Picks are derived from each critic's own scores. Humans vote on matchups on the duels page.