Python API Overview

High-level Python API for programmatic access to conda functionality

Beta Status - While in beta, expect both major and minor changes across minor releases.

Introduction

The conda Python API provides high-level programmatic access to conda’s core functionality. This API allows developers to integrate conda’s package management, solving, and environment management capabilities directly into Python applications.

The API is designed to be more stable and user-friendly than using conda’s internal modules directly.

When to Use the Python API vs CLI

Use the Python API when:

• You need to integrate conda functionality into a Python application
• You want programmatic control over environment solving and package management
• You’re building tools or services that automate conda operations
• You need to query package metadata, caches, or repodata programmatically

Use the CLI when:

• You’re performing interactive or one-off operations
• You’re writing shell scripts or batch files
• You need features not yet exposed in the Python API
• You want the most stable interface (the CLI is considered more stable than the API)

Installation and Imports

The Python API is included with conda. Simply import the classes you need:

from conda.api import Solver, SubdirData, PackageCacheData, PrefixData
from conda.api import DepsModifier, UpdateModifier
from conda.models.match_spec import MatchSpec
from conda.models.channel import Channel

Available APIs

The conda Python API provides the following main classes:

• Solver - High-level access to conda’s dependency solving logic
• SubdirData - Management and querying of repodata.json for subdirectories
• PackageCacheData - Management and querying of package caches
• PrefixData - Management and querying of conda environment prefixes

Quick Example

from conda.api import Solver, SubdirData
from conda.models.match_spec import MatchSpec

# Query available packages
results = SubdirData.query_all("numpy>=1.20")
for record in results:
    print(f"{record.name} {record.version} from {record.channel}")

# Solve for an environment
solver = Solver(
    prefix="/path/to/env",
    channels=["conda-forge", "defaults"],
    subdirs=["linux-64", "noarch"],
    specs_to_add=[MatchSpec("numpy>=1.20"), MatchSpec("pandas")]
)

final_state = solver.solve_final_state()
for package in final_state:
    print(f"{package.name} {package.version}")

Beta Status Notice

All classes and methods in the conda Python API are currently in beta status. This means:

• Both major and minor changes may occur across minor releases
• Method signatures may change
• New parameters may be added or existing ones modified
• Return types may evolve

For production use, pin your conda version and test thoroughly before upgrading.