agentboards.org
Compare/Crush vs DSCode

CrushvsDSCode

Generated from the two spec rows. Green marks the better value where a spec has a clear direction. Everything else is just different.

Crush
Charm · Terminal agent
#25OSSMCP
Panel
6.5
5 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.”

DSCode
ThinkAny · Terminal agent
#153OSS
Panel
6.5
2 spec wins
Reliability
6.2
Usefulness
6.3
Cost
8.0
Longevity
5.7

“It runs agents in parallel, so you can now be behind on four branches at once instead of one.”

Spec by spec

SpecCrushDSCode
Architecture
CategoryTerminal agentTerminal agent
Runslocallocal
Platformsmacos, linux, windowsmacos, linux, windows
Context windownot documentednot documented
Protocols
MCP clientYesNo
MCP serverNoNo
Capabilities
Runs terminal commandsYesYes
Multi-file editsYesYes
Git operationsYesNo
Browser controlNoThe built-in tools stop at web_fetch, web_search and download; there is no browser, CDP or Playwright tool. No
Sandboxed executionNoShell and edit calls are guarded by permission prompts (or skipped with --yolo), with no container or OS-level isolation. YesThe README describes OS sandboxing for the runtime rather than naming a container backend.
Multi-agent orchestrationNoYesParallel agents are listed among the runtime's features.
Headless / CI modeYes`crush run` executes a single non-interactive prompt, reads stdin and supports --quiet for scripts. No
Models
BackboneClaude, GPT, Gemini, any OpenAI-compatible or Anthropic-compatible providerDeepSeek, Codex, OpenAI, Anthropic, OpenRouter, Z.AI, Kimi, MiniMax, xAI, OpenCode Zen
Bring your own modelYesYes
Local modelsYesOllama, LM Studio, llama.cpp and MLX providers are auto-discovered, and any OpenAI- or Anthropic-compatible base_url can be added. No
Cost
Pricing modelbyokbyok
Starts at$0/mo$0/mo
Free tierYesYes
Bring your own keyYesYes
Openness
Open sourceYesYes
LicenseFSL-1.1-MITMIT
GitHub stars28,444367

Which one would each critic pick

CriticCrushDSCodePick
El Juez——not enough reviews
El Amigo7.37.3no preference
El Crítico6.36.3no preference
El Profesor6.56.3Crush — Crush's use of language servers as the context source is the most principled retrieval choice among terminal agents, and its autonomy flag is the most accurately named.
La Inversora5.86.3DSCode — 364 stars and a company name on the row, with a runtime built around one provider's cheapest model: the pitch is price, and price is the easiest thing to undercut.
La Jefa5.86.3DSCode — A signed and notarized desktop app for all three platforms, plus a global npm install for the terminal: two distribution channels, and only one of them I can manage.
El Hacker7.57.0Crush — Reads my LSPs, finds my llama.cpp server on its own, MCP in a JSON file, a --yolo flag, and a license that keeps me from forking it today; close to great, grudgingly.

Picks are derived from each critic's own scores. Humans vote on matchups on the duels page.

Want a third column? The compare tool handles any two agents. Three-way comparisons are on the roadmap once the spec rows are all verified.