accomplish-ai/accomplish

★ 10,788TypeScriptAudience · generalComplexity · 2/5LicenseSetup · easy

Mindmap

mindmap
  root((repo))
    What It Does
      Automates file tasks
      Drafts documents
      Runs browser tasks
      Saves custom workflows
    Privacy
      Data stays on device
      Bring your own API key
      Works fully offline
    AI Providers
      Anthropic OpenAI
      Google DeepSeek
      Local via Ollama
    Platforms
      macOS
      Windows 11
      Ubuntu Linux

mindmap root((repo)) What It Does Automates file tasks Drafts documents Runs browser tasks Saves custom workflows Privacy Data stays on device Bring your own API key Works fully offline AI Providers Anthropic OpenAI Google DeepSeek Local via Ollama Platforms macOS Windows 11 Ubuntu Linux

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Sort and organize a messy downloads folder by file type by typing a single plain-English instruction to the app.

USE CASE 2

Draft a summary of meeting notes or rewrite a document in a different tone without uploading files to a web service.

USE CASE 3

Research a topic in a browser and have the AI compile the results into a document automatically.

USE CASE 4

Define a reusable custom workflow for a repetitive task so the app can repeat it on demand with one command.

Tech stack

TypeScriptElectronReactOllama

Getting it running

Difficulty · easy Time to first run · 30min

Requires an API key from a supported AI provider unless using Ollama for fully local offline operation.

MIT license, use freely for any purpose including commercial projects, just keep the copyright notice in place.

In plain English

Accomplish is a desktop application that lets an AI assistant take real actions on your computer: organizing files, writing documents, and running browser tasks. You download it, give it access to specific folders you choose, and then type instructions in plain English. It does the work instead of just giving you advice. The key distinction from web-based AI tools is that everything stays on your own machine. Your files never leave your device, and Accomplish itself never sees your data. You connect your own API key from an AI provider you already use, such as OpenAI or Anthropic, or you can run it entirely offline using local model software like Ollama with no API key at all. There is no subscription to Accomplish and no data being sent to the company behind it. In practice, you can tell it to sort a messy downloads folder by file type, draft a summary of meeting notes, rewrite a document in a different tone, or research something in a browser and compile the results. You can also define custom workflows you want to reuse, saving them as skills the app can repeat on demand. Every action the AI takes is shown to you first, so you stay in control and can stop it at any point. The app works on macOS, Windows 11, and Ubuntu Linux. It supports a wide range of AI providers including Anthropic, OpenAI, Google AI, DeepSeek, Amazon Bedrock, Azure, and others, as well as local model tools like LM Studio. The source code is publicly available on GitHub under the MIT license, meaning anyone can read it, modify it, or build on it. This project suits people who want an AI assistant that automates repetitive computer tasks without handing their files over to a third-party service. It is aimed at individuals who are comfortable installing a desktop app and connecting an API key, but do not need to write any code to use it.

Copy-paste prompts

Prompt 1

I'm setting up accomplish-ai/accomplish with my Anthropic API key. Walk me through connecting the key and running my first file-organization task on a folder of mixed downloads.

Prompt 2

I want to use accomplish-ai/accomplish fully offline with Ollama and a local model. Which model works best, and how do I point the app to my local Ollama instance?

Prompt 3

Help me define a custom skill in accomplish-ai/accomplish that takes a folder of meeting transcripts, summarizes each one, and saves the summaries as separate text files in a new folder.

Prompt 4

Accomplish previewed an action I want to cancel partway through on Windows 11. How do I review individual steps before they execute and stop the task if I change my mind?

Prompt 5

I want to use accomplish-ai/accomplish with Amazon Bedrock instead of OpenAI. Walk me through selecting Bedrock as the provider and entering the correct credentials in the settings.

Open on GitHub → Explain another repo

← accomplish-ai on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.