Mindket

TinyCodeAgent

Newversion 1 · updated

Overview

An agent that solves tasks by writing Python, running it and fixing its own errors, built on TinyAgent with a Qwen model. Also works as an MCP server. For agent builders and developers.

TinyCodeAgent is a kit in the Developer category, listed by Mindket. It is free. This is version 1, last updated on . Download it on this page, a zip of its files. Nothing in it runs by itself.

At a glance

You need
  • A browser, if you only use the Space online
  • Python and pip, if you run a copy yourself
  • The source credits Modal Functions for running the Python code and Nebius for the Qwen model
Runs in
  • A web browser, on the Space's own page at Hugging Face
  • An MCP client, since the Space carries an MCP tag
  • A copy on your own computer, using Python on Windows, macOS or Linux
Start with
pip install -U huggingface_hub
hf download Agents-MCP-Hackathon/TinyCodeAgent --repo-type space --local-dir TinyCodeAgent
cd TinyCodeAgent

Platform

WindowsOperating system
macOSOperating system
LinuxOperating system

About

What it is

TinyCodeAgent is an AI agent that solves a task by writing Python, running it and fixing its own errors. It can build helper functions as it goes. It is made on the TinyAgent framework, uses Qwen/Qwen3-235B-A22B, and can serve as an MCP server or as a tool for another agent.

Who it suits

Developers who want to see a code-writing agent that tests and repairs its own snippets, and people who want to extend an agent with hooks.

IndustryWho uses itHow they use itExample
AI developmentAgent buildersStudy an agent that writes and debugs codeWatching it retry after a traceback
Data workAnalystsAsk for quick calculationsAsking for the root mean square error of two lists
OperationsDevelopersUse the agent as a tool for another agentConnecting it through MCP

You can skip it if you want a fixed workflow, since the agent decides its own code and tools.

Where to use it

  • A web browser, on the Space's own page at Hugging Face
  • An MCP client, since the Space carries an MCP tag
  • A copy on your own computer, using Python on Windows, macOS or Linux
Use caseWhoWhat they get back
Solve a task with codeAnyoneAn answer found by running Python the agent wrote
Find popular SpacesHobbyistsA list built from Hugging Face's Spaces
Extend an agentDevelopersA base to add hooks, storage and extra MCP connections

How to set it up

  • A browser, if you only use the Space online
  • Python and pip, if you run a copy yourself
  • The source credits Modal Functions for running the Python code and Nebius for the Qwen model
  1. Open the Space's page on Hugging Face to use it online. Nothing needs installing.
  2. For a local copy, install the huggingface_hub package and download the Space's files.
  3. Move into the new folder, then install Gradio and the packages in requirements.txt.
  4. Start the app with python.
pip install -U huggingface_hub
hf download Agents-MCP-Hackathon/TinyCodeAgent --repo-type space --local-dir TinyCodeAgent
cd TinyCodeAgent
pip install gradio==5.33.1
pip install -r requirements.txt

How to use it

Type a task into the chat and let the agent work. It writes code, runs it, reads any error and tries again until the task is solved. On your own computer, start the app first and open the page it serves.

python app.py
ToolWhat it does
python app.pyStarts the agent's web page on your own computer
TinyAgent hooksLet you extend the agent's behavior
Gradio integrationLets another agent use this one as a tool

Example

You ask, "Which Spaces are trending on Hugging Face right now?" The agent writes Python to fetch and list them, and you see the answer it produces.

For newcomers and experts

If you are new to this, try one of the sample questions, such as the root mean square error of two lists. Watch how the agent writes code to answer it.

For experts, the agent keeps state across code runs through Modal Functions rather than a sandbox, which the source says is faster and cheaper. It logs, stores data, supports several MCP connections and reaches other models through LiteLLM. Its system prompt builds on the SmolAgents coding prompt.

  1. You describe a task
  2. The agent writes Python
  3. The code runs and is tested
  4. Errors are read and fixed
  5. The solved result comes back

Good to know

  • Built with Gradio 5.33.1, and the main file is app.py
  • License: apache-2
  • Cost to run: none named, though it relies on Modal Functions and a Qwen model hosted by Nebius
  • Data it reaches: the code it writes and the services it calls, such as Hugging Face's Spaces list
  • Some examples use dummy functions for demonstration, such as the weather and traffic ones

README and files

README

TinyCodeAgent

An agent that solves tasks by writing Python, running it and fixing its own errors, built on TinyAgent with a Qwen model. Also works as an MCP server. For agent builders and developers.

Made by Agents-MCP-Hackathon on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: TinyCodeAgent is an AI agent that solves a task by writing Python, running it and fixing its own errors. It can build helper functions as it goes. It is made on the TinyAgent framework, uses Qwen/Qwen3-235B-A22B, and can serve as an MCP server or as a tool for another agent.

Who it suits: Developers who want to see a code-writing agent that tests and repairs its own snippets, and people who want to extend an agent with hooks.

Industry | Who uses it | How they use it | Example
AI development | Agent builders | Study an agent that writes and debugs code | Watching it retry after a traceback
Data work | Analysts | Ask for quick calculations | Asking for the root mean square error of two lists
Operations | Developers | Use the agent as a tool for another agent | Connecting it through MCP

You can skip it if you want a fixed workflow, since the agent decides its own code and tools.

Where to use it:

  • A web browser, on the Space's own page at Hugging Face
  • An MCP client, since the Space carries an MCP tag
  • A copy on your own computer, using Python on Windows, macOS or Linux
Use case | Who | What they get back
Solve a task with code | Anyone | An answer found by running Python the agent wrote
Find popular Spaces | Hobbyists | A list built from Hugging Face's Spaces
Extend an agent | Developers | A base to add hooks, storage and extra MCP connections

How to set it up:

  • A browser, if you only use the Space online
  • Python and pip, if you run a copy yourself
  • The source credits Modal Functions for running the Python code and Nebius for the Qwen model
  1. Open the Space's page on Hugging Face to use it online. Nothing needs installing.
  2. For a local copy, install the huggingface_hub package and download the Space's files.
  3. Move into the new folder, then install Gradio and the packages in requirements.txt.
  4. Start the app with python.
pip install -U huggingface_hub
hf download Agents-MCP-Hackathon/TinyCodeAgent --repo-type space --local-dir TinyCodeAgent
cd TinyCodeAgent
pip install gradio==5.33.1
pip install -r requirements.txt

How to use it: Type a task into the chat and let the agent work. It writes code, runs it, reads any error and tries again until the task is solved. On your own computer, start the app first and open the page it serves.

python app.py
Tool | What it does
python app.py | Starts the agent's web page on your own computer
TinyAgent hooks | Let you extend the agent's behavior
Gradio integration | Lets another agent use this one as a tool

Example: You ask, "Which Spaces are trending on Hugging Face right now?" The agent writes Python to fetch and list them, and you see the answer it produces.

For newcomers and experts: If you are new to this, try one of the sample questions, such as the root mean square error of two lists. Watch how the agent writes code to answer it.

For experts, the agent keeps state across code runs through Modal Functions rather than a sandbox, which the source says is faster and cheaper. It logs, stores data, supports several MCP connections and reaches other models through LiteLLM. Its system prompt builds on the SmolAgents coding prompt.

You describe a task
The agent writes Python
The code runs and is tested
Errors are read and fixed
The solved result comes back

Good to know:

  • Built with Gradio 5.33.1, and the main file is app.py
  • License: apache-2
  • Cost to run: none named, though it relies on Modal Functions and a Qwen model hosted by Nebius
  • Data it reaches: the code it writes and the services it calls, such as Hugging Face's Spaces list
  • Some examples use dummy functions for demonstration, such as the weather and traffic ones

On Hugging Face, 13 November 2026

  • Likes: 29
  • Built with: gradio
  • License: apache-2.0

Links

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.

Files
README.md  4 KB

1 file, 4 KB, downloaded as a zip.

What’s new

  • Version 1 · 2026-11-12

Compare with similar

Listed Hugging Face Spaces in the same category that share the most tags with this one.

ProductLikesBuilt withLicenseLast change
TinyCodeAgent This page29gradioapache-2.0
Z Image TurboShares mcp-server · gradio1,940gradioapache-2.0
TRELLISShares mcp-server · gradio817gradiomit
SAM3 DemoShares mcp-server · gradio91gradioapache-2.0
TRELLIS.2-Text-to-3D (Rerun)Shares mcp-server · gradio58gradioapache-2.0
Multi-Agent AI - CodingShares mcp-server · gradio39gradioapache-2.0
Numbers as Mindket last read them, on 2026-11-12.

Reviews

No ratings yet

Write a review

No reviews yet.

More from Mindket

See all from Mindket

anaksor-ocr

Sokheng · Kit

Reads Khmer and English text from documents, receipts and photos of signs, using a detector for text lines and a recogniser for each line. Suits developers who…

New

Qwen2.5-7B-Instruct as GGUF files in eight quantization sizes, to chat with locally through llama.cpp. For people who want a capable 7.6B model on their own co…

New

ACE-Step 1.5 XL SFT

ACE-Step · Kit

Generates music from text with a 4B diffusion decoder, 50 steps and guidance control, aiming for higher audio quality than the base. For creators with a GPU of…

New

Similar in Developer

See all in Developer

Claude Code Skills & Plugins

alirezarezvani · Skill

Adds 411 ready-made skills for Claude Code and other coding agents, grouped by role from engineering and product to marketing, finance and compliance. Installs…

New29.8k

buzz

block · Kit

Self-hostable workspace where people and AI agents share the same rooms, built on a Nostr relay that logs every action as a signed event. For teams that want t…

New37.3k

plannotator

backnotprop · Kit

Opens your AI coding agent's plans, specs and code changes in a browser so you can annotate or comment on them and send the feedback straight back to the agent.

New10.5k