repository·main·Indexed 20 days ago
https://github.com/mistralai/client-pythonOfficial Python SDK for accessing Mistral AI's Chat Completion, Embeddings, and Agents APIs. Version 2.8.0 supports multiple cloud providers, including Azure and Google Cloud (Vertex AI), and provides tools for managing agent entities, versioning, and library access levels.
process, which allows users to perform OCR tasks on supported documents.The Workflows.Runs module provides operations to manage and inspect execution runs within Mistral workflows. The available operations are:
list_runs: Retrieve a list of runs.get_run: Retrieve details for a specific run.get_run_history: Retrieve the history associated with a run.mistralgcp SDK is the Python client designed for accessing Mistral AI services hosted on Google Cloud Platform (GCP). It provides programmatic access to Mistral AI's Chat Completion and Embeddings APIs within the GCP ecosystem.The Mistral Python SDK provides access to a wide range of resources and operations, including core AI capabilities, beta features, and observability tools.
complete, stream for chat completions.complete, stream for agentic workflows.complete), Transcriptions (complete, stream), and Voice management (list, create, delete, update, get, get_sample_audio).create embeddings.moderate, moderate_chat) and Classification (classify, classify_chat).process documents.complete, stream).upload, list, retrieve, delete, download, get_signed_url.list, retrieve, delete, update, archive, unarchive).list, create, get, delete, cancel).start, list, get, delete, append, get_history, get_messages, restart, and streaming variants).create, list, get_auth_url, call_tool, etc.).list, create, upload, get_signed_url, etc.).list, create, get, delete, update_metadata, etc.).list, create, get, delete, etc.).execute_workflow, manage deployments, handle executions, and schedule workflows.mistralazure SDK is the Python client for accessing Mistral AI services hosted on Azure. It provides programmatic access to Mistral AI's Chat Completion and Embeddings APIs through the Azure infrastructure. To use these services, you must create an account on La Plateforme to obtain credentials and follow the official Mistral documentation for API specifications.The following aspects of the SDK are unaffected by the v2.x migration:
chat.complete, chat.stream, embeddings.create, fim.complete, files.upload, models.list, fine_tuning.jobs.create) remain the same.>=3.10.pip install mistralai.The Beta Agents API provides the following operations for managing agent entities and their versions:
create: Create a new agent.list_pages: List agent entities using cursor pagination (Recommended).get: Retrieve a specific agent entity.update: Update an existing agent entity.delete: Delete an agent entity.update_version: Update an agent version.list_versions: List all versions of an agent.get_version: Retrieve a specific version of an agent.create_version_alias: Create or update an agent version alias.list_version_aliases: List all aliases for an agent.delete_version_alias: Delete an agent version alias.The Chat Completion API allows you to interact with Mistral's chat models. It supports two primary modes of operation:
complete: A standard request that returns the full chat completion response at once.stream: A streaming request that returns the completion in chunks, allowing for real-time text generation.Refer to the specific operation documentation for detailed parameter requirements and usage examples.
The mistral.beta.libraries.documents namespace provides a set of operations to manage documents within a specific library. This includes listing documents, uploading new ones, retrieving metadata, updating metadata, deleting documents, and accessing content via signed URLs or text extraction.
Note: This is a beta feature and the API surface may change.
complete, which generates a response based on a provided conversation history or prompt. Responses are returned as ChatCompletionResponse objects.beta.libraries API allows you to create and manage libraries, which are used to index documents to enhance agent capabilities. This is currently a beta feature.