servers VS smolagents

Most agents don't speak raw HTTP — they speak MCP. So the real distribution surface is a tiny MCP server that exposes each endpoint as a tool and does the paying under the hood. The agent calls web_search(query); the server hits the paid endpoint, handles the 402 with its operator's wallet, returns JSON. One line to install:

The Model Context Protocol (MCP) is an open standard for connecting AI hosts (Claude, ChatGPT, Cursor, VS Code, and others) to external context and actions through a structured protocol instead of ad-hoc plugins.

Model Context Protocol specification

That changed when I built mpesa-mcp — a Model Context Protocol server that wraps the Daraja API and makes it directly available to Claude, GPT-4, and Gemini.

If you haven't run into it yet, MCP (Model Context Protocol) is an open standard for exposing tools to AI agents. two-go ships an MCP server that runs over stdio with no dependencies, no URL, no account, and no API key. It's all local.

MCP server surface. The CLI is already agent-friendly, so I'm double minded about whether a Model Context Protocol server would make integration smoother. I'm personally not a fan of MCP and prefer skills and CLIs.

As noted in the Model Context Protocol (MCP) Introduction, this protocol allows Claude to interact with systems like GitHub and local databases through a standardized interface.

Keep the project's living context: current task state, decisions, conventions, known footguns as plain Markdown committed to the repo, and expose it to the agent over MCP.

The MCP ecosystem doubles every few months. Bookmark the official MCP servers repository for the canonical list.

HuggingFace built SmolAgents — code-first agents

Something like https://github.com/huggingface/smolagents

Needs a sandbox, otherwise blindly executing generated code is not acceptable

smolagents by huggingface would be more of an agent framework. If it was discussed we would see smolagents/llamaindex/pydantic/etc with frameworks on figure 2. Several frameworks were left off in this paper as it focuses more on the protocols.

https://huggingface.co/docs/smolagents/

The idea was to wrap into an MCP Server the custom-made, sandboxed, local Python interpreter provided with HuggingFace's smolagents library.

I’m curious how this compares to the open-source version made by HuggingFace [1]. As I can tell, the HF version uses season LLMs to search/traverse and parse the web and gather results, then evaluates the results before eventually synthesizing a result.

This version appears to show off a vector store for documents generated from a web crawl (the writer is a vector-store-aaS company)

[1] https://github.com/huggingface/smolagents/tree/main/examples...

Step 1 included python code with matplotlib, which failed because matplotlib was not in the allowed imports. The system then pivoted to another python script which printed a bar plot with ## characters.

Final answer: The waiter with the second most total tips is Corey Johnson

Bar Chart of Total Tips for Each Waiter:

Michael Watts ################################################## (5.67)

Corey Johnson ########## (1.2)

Margaret James ######## (1.0)

[0] - https://github.com/huggingface/smolagents/blob/main/docs/sou...