CLIOvszot
Generated from the two spec rows. Green marks the better value where a spec has a clear direction. Everything else is just different.
CLIO
Synthetic Autonomic Mind · Terminal agent
#81OSSMCP
Panel
6.64 spec wins
- Reliability
- 6.5
- Usefulness
- 6.5
- Cost
- 8.5
- Longevity
- 4.8
“Sub-agent output goes to tmux, GNU Screen or Zellij panes, so the multiplexer argument now has actual stakes.”
zot
Patrice Eckhart · Terminal agent
#110OSS
Panel
6.62 spec wins
- Reliability
- 5.5
- Usefulness
- 6.5
- Cost
- 8.7
- Longevity
- 5.7
“It can run as a Telegram bot, so your coding agent now lives in the same app as your family group chat.”
Spec by spec
| Spec | CLIO | zot |
|---|---|---|
| Architecture | ||
| Category | Terminal agent | Terminal agent |
| Runs | local | local |
| Platforms | macos, linux | macos, linux, windows |
| Context window | not documented | not documented |
| Protocols | ||
| MCP client | Yes | No |
| MCP server | No | No |
| Capabilities | ||
| Runs terminal commands | Yes | Yes |
| Multi-file edits | Yes | Yes |
| Git operations | Yes | No |
| Browser control | YesA web capability fetches and analyses web content; there is no driven browser. | No |
| Sandboxed execution | YesA clio-container mode provides OS-level isolation via Docker; CLIO itself also runs from a published container image. | NoThe built-in /jail sandbox is a path and shell guardrail, not container isolation; the README suggests running zot under Docker if you need real isolation. |
| Multi-agent orchestration | YesSub-agents run in parallel with file locks, git locks and rate limiting, with live output panes in tmux, GNU Screen or Zellij. | YesSwarm subagents share the host working directory and the same read, write, edit and bash tools; there is no per-agent worktree or branch. |
| Headless / CI mode | No | YesPrint, stream and JSON modes accept piped stdin, and a headless Telegram daemon is supported. |
| Models | ||
| Backbone | GitHub Copilot, Anthropic, OpenAI, Google Gemini, DeepSeek, OpenRouter, Z.AI, NVIDIA NIM, llama.cpp, LM Studio | Anthropic, OpenAI, Codex, Google Gemini, Vertex AI, GitHub Copilot, Bedrock, Azure OpenAI, OpenRouter, Groq, Cerebras, xAI, Together, Hugging Face, Mistral, Moonshot, Kimi, DeepSeek, Z.AI, Xiaomi, MiniMax, Fireworks, Vercel AI Gateway, Cloudflare AI, Ollama |
| Bring your own model | Yes | Yes |
| Local models | Yes | YesOllama and local models are a built-in provider. |
| 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 | GPL-3.0 | MIT |
| GitHub stars | 231 | 348 |
Which one would each critic pick
| Critic | CLIO | zot | Pick |
|---|---|---|---|
| El Juez | — | — | not enough reviews |
| El Amigo | 7.0 | 7.0 | no preference |
| El Crítico | 6.0 | 6.3 | zot — The /jail sandbox blocks obvious shell escapes and the README's own advice is to run the whole thing under Docker if you need real isolation. |
| El Profesor | 7.3 | 6.3 | CLIO — Approval sits between the plan and the edits, and any turn can be undone, so the loop has both a gate before it and a reverse after it. |
| La Inversora | 5.8 | 6.0 | zot — Thirty-plus provider integrations maintained by one person is a liability rather than an asset: every one is an API that changes without asking. |
| La Jefa | 5.5 | 6.3 | zot — The print and JSON modes accept piped input and emit structured output, so it can sit in a pipeline, and Windows is supported where most of this category is not. |
| El Hacker | 8.0 | 7.8 | CLIO — GPL-3.0, MCP servers attach, skills load from git repositories, and llama.cpp or LM Studio means the weights can be the ones on my machine. |
Picks are derived from each critic's own scores. Humans vote on matchups on the duels page.