Agent framework - verified
smolagents (Hugging Face): features, install and latest version
smolagents (Hugging Face) is an AI agent framework maintained by Hugging Face. As its docs describe it: smolagents is an open-source Python library from Hugging Face for building and running agents in a few lines of code, with first-class support for CodeAgent, which writes its actions as Python code, alongside a conventional ToolCallingAgent. It supports Python. License: Apache-2.0. Latest release: Python package smolagents 1.26.0 (May 29, 2026).
- Languages
- 1
- Latest Python
- 1.26.0
- Released
- 2026-05-29
- Features documented
- 6 of 10
Install
Copied verbatim from the page linked under each command.
Python
pip install smolagentsQuickstart
The first example from the official docs, unchanged. It may need an API key for the model provider it calls.
from smolagents import CodeAgent, InferenceClientModel
# Initialize a model (using Hugging Face Inference API)
model = InferenceClientModel() # Uses a default model
# Create an agent with no tools
agent = CodeAgent(tools=[], model=model)
# Run the agent with a task
result = agent.run("Calculate the sum of numbers from 1 to 10")
print(result)Features
What the smolagents (Hugging Face) docs say it supports. Not stated means the docs do not say either way.
| Feature | Value | Source |
|---|---|---|
| MCP | Yes | huggingface.co/docs/smolagents/index |
| Multi-agent | Yes | huggingface.co/docs/smolagents/guided_tour |
| Durable execution | Not stated | Docs do not say |
| Human in the loop | Yes | huggingface.co/docs/smolagents/examples/plan_customization |
| Streaming | Yes | huggingface.co/docs/smolagents/reference/agents |
| Structured output | Not stated | Docs do not say |
| Tracing | Yes | huggingface.co/docs/smolagents/tutorials/inspect_runs |
| Memory | Not stated | Docs do not say |
| Evals | Not stated | Docs do not say |
| Model-agnostic | Yes | huggingface.co/docs/smolagents/index |
| Languages | Python | See sources below |
| License | Apache-2.0 | github.com/huggingface/smolagents/blob/main/LICENSE |
| Latest Python release | smolagents 1.26.0May 29, 2026 | pypi.org/project/smolagents |
Compare smolagents (Hugging Face)
FAQ
Does smolagents (Hugging Face) support MCP?
Which languages does smolagents (Hugging Face) support?
Python. Published packages: smolagents (Python).
How do I install smolagents (Hugging Face)?
For Python: "pip install smolagents", as shown at https://huggingface.co/docs/smolagents/installation. This page lists 1 install command, each copied from the linked source.
What is the latest version of smolagents (Hugging Face)?
As of October 18, 2026: Python package smolagents 1.26.0 (May 29, 2026), https://pypi.org/project/smolagents/.
What license is smolagents (Hugging Face) released under?
Apache-2.0. The license is at https://github.com/huggingface/smolagents/blob/main/LICENSE.
Does smolagents (Hugging Face) support durable execution?
Its official docs, read on October 18, 2026, do not say.
Does smolagents (Hugging Face) support human-in-the-loop?
Yes, per https://huggingface.co/docs/smolagents/examples/plan_customization
Is smolagents (Hugging Face) model-agnostic?
Which features does smolagents (Hugging Face) document?
6 of the 10 tracked here: MCP, Multi-agent, Human in the loop, Streaming, Tracing and Model-agnostic. Not stated in its docs: Durable execution, Structured output, Memory and Evals.
Sources
Last verified from the pages below.
- Docs home and quickstart - huggingface.co
- Installation - huggingface.co
- Guided tour (multi-agents) - huggingface.co
- Agents reference - huggingface.co
- Inspecting runs with OpenTelemetry - huggingface.co
- Manage your agent's memory (in-run step memory) - huggingface.co
- Human-in-the-Loop plan customization - huggingface.co
- GitHub repo - github.com
- PyPI package - pypi.org
Newsletter
Agent frameworks, every week
One email per week. Framework releases, new agent tooling and what they mean for what you build.