mcp-typescript-template vs penguin-harness
Two of the top projects in Agent and MCP starters, side by side: who leads on each score part and who each one suits, then setup, license, activity and what each review found.
mcp-typescript-template
TypeScript starter for remote MCP servers over HTTP with Express and Effect
penguin-harness
Agent framework that evolves an agent's prompts, skills and tools against benchmarks
Verdict
mcp-typescript-template ranks higher, 63 to 61. Of the 5 score parts we measure for both, mcp-typescript-template leads on 3, penguin-harness on 1 and 1 is even.
- Adoptionpenguin-harness leads by 49 points, 49 to 0: a wide gap.
- FreshnessEven at 100.
- Maintenancemcp-typescript-template leads by 4 points, 97 to 93: close to even.
- Easy to runmcp-typescript-template leads by 17 points, 50 to 33: a wide gap.
- Agent-readymcp-typescript-template leads by 85 points, 85 to 0: a wide gap.
Pick mcp-typescript-template if you are
Developers starting a remote MCP server in TypeScript
- Example tools cover structured output, annotations and MCP elicitation with accept/decline/cancel handling
- Stateless per-request server following the MCP 2026-07-28 spec, with fallback for older clients
Pick penguin-harness if you are
Engineers tuning agent prompts and skills, or self-hosting agents
- Five RSI algorithms run on the same benchmarks, scoreboard and per-version snapshots
- Desktop app, Web App, CLI and SDK share one engine and the ~/.penguin/data directory
| What we compare | mcp-typescript-template | penguin-harness |
|---|---|---|
| Score parts, out of 100 | ||
| Adoption | 0, niche | 49, known |
| Freshness | 100, active | 100, active |
| Maintenance | 97, healthy | 93, healthy |
| Easy to run | 50, easy | 33, some setup |
| Agent-ready | 85, ready | 0, none |
| Facts from GitHub and the README | ||
| Stars | 58 | 2.9k |
| License | MIT (permissive) | Apache-2.0 (permissive) |
| Last commit | Oct 2026 | Oct 2026 |
| Last release | Oct 2026 | Oct 2026 |
| Language | TypeScript | TypeScript |
| Docker | Yes | Yes |
| GPU | Not needed | Not needed |
| arm64 or Apple Silicon | Not stated | Mentioned |
mcp-typescript-template
A degit-able template for building stateless, remote MCP servers on Node 24.11+ using the MCP TypeScript SDK, Express and Effect. It ships two example tools (echo and elicit_echo) with typed input and output schemas, plus an example Agent Skill published through the MCP Skills extension. Build, test, lint and format run through the Vite+ vp CLI.
Who it is for: Developers starting a remote MCP server in TypeScript
Strengths
- Example tools cover structured output, annotations and MCP elicitation with accept/decline/cancel handling
- Stateless per-request server following the MCP 2026-07-28 spec, with fallback for older clients
- Includes Dockerfile, compose file with health check, and in-memory client/server integration tests
- Bundles an example Agent Skill served via skills/list, skills/get and resources/read
Weaknesses
- Requires Node.js 24.11 or newer; Vite+ does not run on earlier Node 24 releases
- Tool code is written in Effect, which adds a learning curve for teams not using it
- HTTP transport only; the README mentions no stdio transport
- No auth or database included; only two toy example tools
- TypeScript
- Needs Node.js 24.11+, Express, Effect, @modelcontextprotocol/server, Vite+
- GitHub template
- Docker
penguin-harness
PenguinHarness runs agents with a small toolset and keeps the model fixed while it evolves the harness (prompts, Skills, tools, hooks) over scored rounds. Five RSI algorithms (default, OPRO, APE, ACE, AWM) ship as plugins and run on built-in Harbor benchmarks in Docker, with each candidate snapshotted as a version. It ships as a desktop app, Web App, CLI and TypeScript SDK, and can also generate agent apps from a prompt.
Who it is for: Engineers tuning agent prompts and skills, or self-hosting agents
Strengths
- Five RSI algorithms run on the same benchmarks, scoreboard and per-version snapshots
- Desktop app, Web App, CLI and SDK share one engine and the ~/.penguin/data directory
- Works with many providers plus custom OpenAI-protocol endpoints, including local models
- Agents can run on remote machines over ssh and drive Chrome through an extension
Weaknesses
- Benchmarks run as Harbor tasks in Docker, so self-evolution needs Docker
- Cost and accuracy claims come from the project's own suite, not independent runs
- RAM and disk requirements are not documented
- Company-level self-evolution and OpenShell integration are still unchecked roadmap items
- TypeScript, OpenAI, Anthropic, Google Gemini, OpenRouter
- Needs Node >= 24 (npm install), Docker (benchmarks), Model provider API key
- Docker
- env example file
More in Agent and MCP starters
Building with these?
We help teams pick an open-source AI stack and run it in production. Client work never changes a rank.