| Name | autoplan JSON |
| Version |
0.1.2
JSON |
| download |
| home_page | https://Institute-for-Advanced-Enterprise-AI.github.io/autoplan |
| Summary | AutoPlan is an open-source Python framework that provides a powerful pattern for implementing agentic AI applications, leveraging dynamic plan generation to select and use external tools based on the task's context. |
| upload_time | 2025-02-20 02:11:40 |
| maintainer | None |
| docs_url | None |
| author | Dave Nachman |
| requires_python | <4.0,>=3.12 |
| license | MIT |
| keywords |
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# AutoPlan
AutoPlan is an open-source Python framework that provides a powerful pattern for implementing agentic AI applications, leveraging dynamic plan generation to select and use external tools based on the task's context.
## Installation
AutoPlan requires Python 3.12 or higher.
Install using pip:
```
pip install -U autoplan
```
## Quick start
The best way to start using AutoPlan is to create a new application. AutoPlan has a command-line interface that allows you to do that. To create your own application, you can use the following command:
```bash
autoplan generate \
--name "my_app" \
--description "Given a situation described by the user, generate a joke about it." \
--outdir .
```
Note that you can also leave the parameters empty and let the CLI prompt you for the information, which will help you get started with bootstrapping your application.
This command line will create a new folder named `my_app` with the basic structure of an AutoPlan application. You can then go in the application folder and run it with the following command:
```bash
cd my_app
poetry install
poetry run python my_app/app.py
```
This will start a Gradio interface that allows you to interact with the application without any additional effort. You can now use your browser and go to `http://localhost:7860` to see the application.
> [!NOTE]
> By default AutoPlan will use Open AI models and will include a search tool based on [you.com](http://api.you.com), which require API keys. You can set the `OPENAI_API_KEY` and `YDC_API_KEY` environment variables to your OpenAI and You API keys to use your own accounts.
> [!NOTE]
> You may want to use other LLMs in your application. You can do that by setting the `generate_plan_llm_model` and `combine_steps_llm_model` parameters in the `with_planning` decorator, and/or by setting the model of your choice in your tool implementations. If your application uses other models, don't forget to set the API keys for those models in your environment (e.g. `ANTHROPIC_API_KEY = <your-key>`) .
## Why use AutoPlan?
Agentic AI applications are an emerging AI paradigm where LLMs use external tools to accomplish tasks that are beyond their own capabilities while keeping control over how to use those tools so that their abilities are not limited to predefined workflows. Building agentic applications requires a system design that allows plans to be dynamically generated, tools to be efficiently executed, and data flowing between tools to be properly channeled to provide a coherent output. AutoPlan provides just that.
AutoPlan is organized around three core components, each serving a specific purpose to enable dynamic planning, execution, and integration required for building an agentic application:
**Tools** can be any typed Python function — they can be procedural code, LLM calls, or AutoPlan applications themselves. Tools can be composed from smaller tools.
**Planners** are LLM-based components that generate the sequence of tools to be executed and the arguments to be passed to each tool to solve a given task.
**Composers** integrate tool outputs based on the planner’s strategy to produce a final output.
## Developer docs
### Type checking
For the planning framework: `poetry run pyright autoplan`
### Generating docs
Run a web server to view the docs:
`poetry run pdoc autoplan`
Generate the docs:
`poetry run pdoc autoplan --output build-docs`
### Running unit tests
`poetry run pytest tests`
Raw data
{
"_id": null,
"home_page": "https://Institute-for-Advanced-Enterprise-AI.github.io/autoplan",
"name": "autoplan",
"maintainer": null,
"docs_url": null,
"requires_python": "<4.0,>=3.12",
"maintainer_email": null,
"keywords": null,
"author": "Dave Nachman",
"author_email": "dave.nachman.dev@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/65/77/13bf388fbc55390121112e611d33ac9b6241e03a34dcf9b75e59ff0b8a41/autoplan-0.1.2.tar.gz",
"platform": null,
"description": "# AutoPlan\n\nAutoPlan is an open-source Python framework that provides a powerful pattern for implementing agentic AI applications, leveraging dynamic plan generation to select and use external tools based on the task's context.\n\n\n## Installation\n\nAutoPlan requires Python 3.12 or higher.\n\nInstall using pip:\n\n```\npip install -U autoplan\n```\n\n\n## Quick start\n\nThe best way to start using AutoPlan is to create a new application. AutoPlan has a command-line interface that allows you to do that. To create your own application, you can use the following command:\n\n```bash\nautoplan generate \\\n --name \"my_app\" \\\n --description \"Given a situation described by the user, generate a joke about it.\" \\ \n --outdir .\n```\n\nNote that you can also leave the parameters empty and let the CLI prompt you for the information, which will help you get started with bootstrapping your application.\n\nThis command line will create a new folder named `my_app` with the basic structure of an AutoPlan application. You can then go in the application folder and run it with the following command:\n\n```bash\ncd my_app\npoetry install\npoetry run python my_app/app.py\n```\n\nThis will start a Gradio interface that allows you to interact with the application without any additional effort. You can now use your browser and go to `http://localhost:7860` to see the application.\n\n> [!NOTE]\n> By default AutoPlan will use Open AI models and will include a search tool based on [you.com](http://api.you.com), which require API keys. You can set the `OPENAI_API_KEY` and `YDC_API_KEY` environment variables to your OpenAI and You API keys to use your own accounts. \n\n> [!NOTE]\n> You may want to use other LLMs in your application. You can do that by setting the `generate_plan_llm_model` and `combine_steps_llm_model` parameters in the `with_planning` decorator, and/or by setting the model of your choice in your tool implementations. If your application uses other models, don't forget to set the API keys for those models in your environment (e.g. `ANTHROPIC_API_KEY = <your-key>`) .\n\n## Why use AutoPlan?\n\nAgentic AI applications are an emerging AI paradigm where LLMs use external tools to accomplish tasks that are beyond their own capabilities while keeping control over how to use those tools so that their abilities are not limited to predefined workflows. Building agentic applications requires a system design that allows plans to be dynamically generated, tools to be efficiently executed, and data flowing between tools to be properly channeled to provide a coherent output. AutoPlan provides just that.\n\nAutoPlan is organized around three core components, each serving a specific purpose to enable dynamic planning, execution, and integration required for building an agentic application:\n\n**Tools** can be any typed Python function \u2014 they can be procedural code, LLM calls, or AutoPlan applications themselves. Tools can be composed from smaller tools.\n\n**Planners** are LLM-based components that generate the sequence of tools to be executed and the arguments to be passed to each tool to solve a given task. \n\n**Composers** integrate tool outputs based on the planner\u2019s strategy to produce a final output.\n\n\n\n\n## Developer docs\n\n### Type checking\n\nFor the planning framework: `poetry run pyright autoplan`\n\n### Generating docs\n\nRun a web server to view the docs:\n`poetry run pdoc autoplan`\n\nGenerate the docs:\n`poetry run pdoc autoplan --output build-docs`\n\n### Running unit tests\n\n`poetry run pytest tests`\n\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "AutoPlan is an open-source Python framework that provides a powerful pattern for implementing agentic AI applications, leveraging dynamic plan generation to select and use external tools based on the task's context.",
"version": "0.1.2",
"project_urls": {
"Documentation": "https://Institute-for-Advanced-Enterprise-AI.github.io/autoplan",
"Homepage": "https://Institute-for-Advanced-Enterprise-AI.github.io/autoplan",
"Repository": "https://github.com/Institute-for-Advanced-Enterprise-AI/autoplan"
},
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "d8b794eff0a21c25cbfa37f1645bbd074fa7aa9d980f44e38026d9849a75bbfa",
"md5": "18b58f9fb4426756197cae0e1ab1b0b7",
"sha256": "dbafbca6e55f27a5e5e6a54de9c98548d1539ceb5bad4ca551b8e3624a6ed619"
},
"downloads": -1,
"filename": "autoplan-0.1.2-py3-none-any.whl",
"has_sig": false,
"md5_digest": "18b58f9fb4426756197cae0e1ab1b0b7",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "<4.0,>=3.12",
"size": 25656,
"upload_time": "2025-02-20T02:11:39",
"upload_time_iso_8601": "2025-02-20T02:11:39.927752Z",
"url": "https://files.pythonhosted.org/packages/d8/b7/94eff0a21c25cbfa37f1645bbd074fa7aa9d980f44e38026d9849a75bbfa/autoplan-0.1.2-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "657713bf388fbc55390121112e611d33ac9b6241e03a34dcf9b75e59ff0b8a41",
"md5": "502f0b0b5cb69b1366f1b5671bbd87ea",
"sha256": "bdfbfb6a348a447d6c740134456d515e13921b3fcfcb6fffc0bbd0e17afaa08e"
},
"downloads": -1,
"filename": "autoplan-0.1.2.tar.gz",
"has_sig": false,
"md5_digest": "502f0b0b5cb69b1366f1b5671bbd87ea",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "<4.0,>=3.12",
"size": 19741,
"upload_time": "2025-02-20T02:11:40",
"upload_time_iso_8601": "2025-02-20T02:11:40.999928Z",
"url": "https://files.pythonhosted.org/packages/65/77/13bf388fbc55390121112e611d33ac9b6241e03a34dcf9b75e59ff0b8a41/autoplan-0.1.2.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-02-20 02:11:40",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "Institute-for-Advanced-Enterprise-AI",
"github_project": "autoplan",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "autoplan"
}
