Continue CLIvsCrush
Generated from the two spec rows. Green marks the better value where a spec has a clear direction. Everything else is just different.
Continue CLI
Continue · Terminal agent
#24OSSMCP
Panel
6.41 spec wins
- Reliability
- 6.2
- Usefulness
- 6.3
- Cost
- 8.0
- Longevity
- 5.2
“It installs by piping a URL straight into your shell, then politely asks permission before it touches a single file.”
Crush
Charm · Terminal agent
#25OSSMCP
Panel
6.50 spec wins
- Reliability
- 6.0
- Usefulness
- 6.5
- Cost
- 7.3
- Longevity
- 6.2
“Has a --yolo flag, auto-discovers your local llama.cpp, and a license with MIT in the name that is not MIT yet.”
Spec by spec
| Spec | Continue CLI | Crush |
|---|---|---|
| Architecture | ||
| Category | Terminal agent | Terminal agent |
| Runs | local | local |
| Platforms | macos, linux, windows | macos, linux, windows |
| Context window | not documented | not documented |
| Protocols | ||
| MCP client | Yes | Yes |
| MCP server | No | No |
| Capabilities | ||
| Runs terminal commands | Yes | Yes |
| Multi-file edits | Yes | Yes |
| Git operations | Yes | Yes |
| Browser control | No | NoThe built-in tools stop at web_fetch, web_search and download; there is no browser, CDP or Playwright tool. |
| Sandboxed execution | NoThe docs describe running the CLI inside a Docker container, not sandboxing the agent in one. | NoShell and edit calls are guarded by permission prompts (or skipped with --yolo), with no container or OS-level isolation. |
| Multi-agent orchestration | No | No |
| Headless / CI mode | Yes | Yes`crush run` executes a single non-interactive prompt, reads stdin and supports --quiet for scripts. |
| Models | ||
| Backbone | any provider configured in config.yaml, Anthropic | Claude, GPT, Gemini, any OpenAI-compatible or Anthropic-compatible provider |
| Bring your own model | Yes | Yes |
| Local models | YesThe CLI reads the same ~/.continue/config.yaml provider schema as the IDE extensions, including a custom apiBase, so any local server can be configured; the CLI pages defer to that reference rather than naming Ollama or LM Studio. | YesOllama, LM Studio, llama.cpp and MLX providers are auto-discovered, and any OpenAI- or Anthropic-compatible base_url can be added. |
| 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 | Apache-2.0 | FSL-1.1-MIT |
| GitHub stars | 36,084 | 28,444 |
Which one would each critic pick
| Critic | Continue CLI | Crush | Pick |
|---|---|---|---|
| El Juez | — | — | not enough reviews |
| El Amigo | 6.8 | 7.3 | Crush — Crush is the terminal agent to pick if you want LSP-grade context and a Charm-quality interface with your own keys, and the one to read the license on before you fork it. |
| El Crítico | 6.3 | 6.3 | no preference |
| El Profesor | 6.5 | 6.5 | no preference |
| La Inversora | 5.3 | 5.8 | Crush — Charm put its coding agent under a source-available license and attached a $20 Hyper plan, which says they intend to sell it, and the terminal aesthetic is the distribution. |
| La Jefa | 6.0 | 5.8 | Continue CLI — Zero licence for sixty and a single-shot flag that fits our pipelines, but sign-in goes to a vendor account by default and nothing on the row mentions SSO or audit. |
| El Hacker | 7.8 | 7.5 | Continue CLI — Apache-2.0 with the whole thing in a public monorepo, and a provider block that takes a custom apiBase, which is how my own server gets in. |
Picks are derived from each critic's own scores. Humans vote on matchups on the duels page.