How to install a specific package version with Conda

Dependency handoffs break when a Conda environment takes a newer package than the project was tested against. A version-constrained install tells Conda which package release must satisfy the solve before it changes the target environment.

conda install accepts package specs such as requests=2.32.5 and applies them to either the active environment or a named environment selected with --name. Searching first confirms that the requested version exists in the configured channels before the solver starts a transaction that cannot succeed.

The install constraint affects the current transaction. Future update commands can still move the package unless the environment also has a pin, so record exact versions in the project environment file or package pin policy when reproducibility matters after installation.

1. Open a terminal where Conda is available.
2. Accept a blocked Anaconda channel Terms of Service only after review.

   $ conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/main
   accepted Terms of Service for https://repo.anaconda.com/pkgs/main

   Run this only for channels your organization allows. If Conda lists more than one blocked channel, repeat the command for each listed URL or remove the channel instead.

3. List the environments and choose the target environment.

   $ conda env list
   
   # conda environments:
   #
   # * -> active
   # + -> frozen
   base                     /opt/conda
   analytics                /opt/conda/envs/analytics

   Use an existing environment for the install. Create the environment first if the target name is missing.
   Related: How to list Anaconda environments
   Related: How to create a Conda environment

4. Search for available versions of the package.

   $ conda search --full-name requests
   Loading channels: done
   # Name                       Version           Build  Channel
   requests                      2.24.0            py_0  pkgs/main
   requests                      2.25.0    pyhd3eb1b0_0  pkgs/main
   ##### snipped #####
   requests                      2.32.4 py313hd43f75c_0  pkgs/main
   requests                      2.32.5 py313hd43f75c_0  pkgs/main
   requests                      2.32.5 py313hd43f75c_1  pkgs/main
   requests                      2.33.1 py313hd43f75c_0  pkgs/main
   requests                      2.34.2 py313hd43f75c_0  pkgs/main

   Replace requests with the package name. Use the same channel settings or --channel option that the target environment should use.
   Related: How to search for Conda package versions

5. Install the requested package version into the target environment.

   $ conda install --name analytics "requests=2.32.5" --yes
   Channels:
    - defaults
   Platform: linux-aarch64
   Collecting package metadata (repodata.json): done
   Solving environment: done
   
   ## Package Plan ##
   
     environment location: /opt/conda/envs/analytics
   
     added / updated specs:
       - requests=2.32.5
   
   The following NEW packages will be INSTALLED:
   
     brotlicffi         pkgs/main/linux-aarch64::brotlicffi-1.2.0.0-py313h5b11e9f_0
     certifi            pkgs/main/linux-aarch64::certifi-2026.5.20-py313hd43f75c_0
   ##### snipped #####
     requests           pkgs/main/linux-aarch64::requests-2.32.5-py313hd43f75c_1
     urllib3            pkgs/main/linux-aarch64::urllib3-2.7.0-py313hd43f75c_0
   
   Preparing transaction: done
   Verifying transaction: done
   Executing transaction: done

   Leave off --yes when the package plan should be reviewed before the transaction runs. Use an exact version such as requests=2.32.5 when the project needs that release, or a narrow series such as requests=2.32.* only when patch-level movement is allowed.

6. Confirm the installed version in the environment package list.

   $ conda list --name analytics requests
   # packages in environment at /opt/conda/envs/analytics:
   #
   # Name                     Version          Build            Channel
   requests                   2.32.5           py313hd43f75c_1

   The Version column should match the requested version. If the package name appears more than once, review channel priority and build constraints before using the environment.
   Related: How to list packages in an Anaconda environment

7. Run a package-specific smoke test from the same environment.

   $ conda run --name analytics python -c "import requests; print(requests.__version__)"
   2.32.5

   Replace the sample Python import with the command, notebook import, or executable that proves the installed package works for the project.