Robocorp is a development framework for automation and AI Actions for Python, designed to help developers create, deploy, and run AI Actions and automation tasks, and operate them anywhere. Its repository contains the core Python libraries of the Robocorp automation stack, used for configuring package entry points, controlling logging, and other functions, and provides a large number of libraries for common automation tasks.
Purpose
- Create and run Python AI Actions and automation tasks.
- Deploy tasks and actions to any environment and operate them.
- Provide support for executable actions for AI agents and assistants.
Main Capabilities
- Core features include defining Tasks, logging, Work Items, Vault, Assets, and Control Room integration.
- AI Actions, Data Actions, and Action Server are located in separate PyPI packages: sema4ai-actions, sema4ai-data, sema4ai-action-server.
- Includes framework libraries and common automation libraries; other functionality can be implemented through the Python ecosystem.
- Additional automation capabilities can refer to rpaframework, and commonly used third-party libraries and their code snippets are collected.
- Provides the AI coding agent skill robocorp-automation, which can teach agents how to bootstrap, build, run, and validate Robocorp task packages (using robocorp.tasks and rcc).
Usage
- The simplest way is to install the robocorp metapackage, which includes the core framework functionality.
- Each component can also be installed as a separate dependency: for automation task packages, list them in conda.yaml; for action packages, list them in package.yaml; in other cases, they can be listed in requirements.txt, pyproject.toml, setup.py|cfg, and other files.
- Libraries not included by default in robocorp (without the "Metapackage" marker) need to be manually added as dependencies.
- When using Claude Code, the skill can be installed from this repository as a plugin; other coding agents can copy the skill into the automation project.
Applicable Scenarios
- Build Python automation task packages and AI Actions.
- Automation projects that require task definitions, logging, work items, secrets management, assets, and control room integration.
- Integrate executable actions into AI agents or assistants.
Limitations and Notes
- Versions follow SemVer; 0.x.x versions are in the development stage, and updates may introduce breaking changes; the goal is to reach 1.0 as soon as possible, after which SemVer rules will be followed.
- When introducing breaking changes, a new major version will be released, and the changelog should include a migration guide.
- Modules starting with an underscore _ are not part of the public API and should not be imported directly; only functions and classes under the robocorp.<library> namespace should be used.
- If access to other private objects is needed, a feature request must be created.
- Specific installation commands, code examples, platform support, pricing information, etc. are not provided in the materials and are therefore not provided here.
