Build a Chainlit App with Mistral AI

The goal of this cookbook is to show how one can build a Chainlit application on top of Mistral AI's APIs!

We will highlight the reasoning capabilities of Mistral's LLMs by letting a self-reflective agent assess whether it has gathered enough information to answer nested user questions, such as "What is the weather in Napoleon's hometown?"

To answer such questions, our application should go through multiple-step reasoning: first get Napoleon's hometown, then fetch the weather for that location.

You can read through this notebook or simply go with chainlit run app.py since the whole code is in app.py. You will find here a split of the whole application code with explanations:

• Setup
• Define available tools
• Agent logic
• On message callback
• Starter questions

Here's a visual of what we will have built in a few minutes:

Setup

Requirements

We will install mistralai, chainlit and python-dotenv.

Be sure to create a .env file with the line MISTRAL_API_KEY= followed by your Mistral AI API key.

Optional - Tracing

You can get a LITERAL_API_KEY from Literal AI to setup tracing and visualize the flow of your application.

Within the code, Chainlit offers the @chainlit.step decorators to trace your functions, along with an automatic instrumentation of Mistral's API via chainlit.instrument_mistralai().

The trace for this notebook example is: https://cloud.getliteral.ai/thread/ea173d7d-a53f-4eaf-a451-82090b07e6ff.

Define available tools

In the next cell, we define the tools, and their JSON definitions, which we will provide to the agent. We have two tools:

• get_current_weather -> takes in a location
• get_home_town -> takes in a person's name

Optionally, you can decorate your tool definitions with @cl.step(), specifying a type and name to organize the traces you can visualize from Literal AI.

Agent logic

For the agent logic, we simply repeat the following pattern (max. 5 times):

• ask the user question to Mistral, making both tools available
• execute tools if Mistral asks for it, otherwise return message

You will notice that we added an optional @cl.step of type run and with optional tags to trace the call accordingly in Literal AI.

Visual trace: https://cloud.getliteral.ai/thread/ea173d7d-a53f-4eaf-a451-82090b07e6ff

On message callback

The callback below, properly annotated with @cl.on_message, ensures our run_agent function is called upon every new user message.

Starter questions

You can define starter questions for your users to easily try out your application, which will look like this:

We have got many more Chainlit features in store (authentication, feedback, Slack/Discord integrations, etc.) to let you build custom LLM applications and really take advantage of Mistral's LLM capabilities.

Please visit the Chainlit documentation to learn more!