claude-quickstarts
Starter projects that show how to build apps with the Claude API, from support agents to browser and desktop a
- Stars
- 19.1k
- Forks
- 3.3k
- Licence
- MIT
- Language
- TypeScript
- Last update
- 6 days ago
- Type
- Learning resource
What it is
Claude Quickstarts is a collection of example projects from Anthropic meant to help developers begin building applications on the Claude API. Each project is a starting point you can run, read, and then adapt to your own needs. The collection covers a range of examples, including a customer support agent, a financial data analyst, computer and browser use demos, and several Managed Agents integrations with tools like Slack, Linear, and Sentry.
- Run the customer support agent quickstart to see an AI support assistant that draws on a knowledge base
- Explore the financial data analyst quickstart to see chat-based analysis with interactive charts
- Study the browser use demo to learn how Claude can navigate sites and fill forms through a custom browser tool
- Read the Managed Agents Slack or Linear examples to see how a webhook can drive an agent session
Great for
- Developers who want to see working Claude API examples before writing their own app
- Teams evaluating which Claude capabilities fit a use case such as support, data analysis, or browser automation
- Developers already using Slack, Linear, or Sentry who want an example integration to study
- Students learning how AI agents are wired together with tools and sessions
Not ideal for
- Non-developers who want a ready-made app without writing or configuring code, since the README assumes API setup
- Anyone needing production-ready code, because several projects say they are reference or teaching examples rather than production code
- Users without a Claude API key, since the README says one is needed to use these quickstarts
How to get it
- 1
Clone and build from source
You'll need: Claude API key from console.anthropic.com · Git
Clone the repository
git clone https://github.com/anthropics/claude-quickstarts - 2
Run it
- Sign up at console.anthropic.com and create an API key
- Clone the repository with git
- Open the folder of the quickstart you want, such as customer-support-agent
- Follow that folder's own README to install dependencies and set your API key
- Run the project and try a sample request
- 3
Get the most out of it
- Start with the customer support agent, which the README describes as a foundation you can customize
- Read the Computer Toolset and Browser Toolset quickstarts to see the minimal interface before the full demos
- Use the Managed Agents Daily Brief example to see how a scheduled deployment is defined as files
- Check the Autonomous Coding Agent to see how progress is persisted across sessions via git
Downloads
Strengths
- Maintained by Anthropic and published under the MIT license, so the code can be reused and modified
- Covers many different use cases, from chat apps to scheduled agents and browser or desktop control
- Each quickstart is in its own folder, so you can work on one example without setting up the others
Trade-offs
- Projects vary a lot in complexity and maturity; some are explicitly minimal examples and not production code
- The README is truncated and does not give full setup steps for each project, so you will need to read each subfolder's instructions
- Several projects depend on specific Claude features, such as Managed Agents or newer toolset versions, which may change over time
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| claude-quickstarts (this) | 19.1k | Learning resource | intermediate | 6 days ago | Developers who want to see working Claude API examples before writing their own app |
| langchain | 149k | Framework | intermediate | 4 days ago | Python developers building LLM-powered applications |
| llama_index | 52.8k | Library | intermediate | 10 days ago | Python developers building retrieval-augmented generation (RAG) apps over private documents |
| openai-cookbook | 76.9k | Learning resource | beginner | 5 days ago | Developers who are new to the OpenAI API and want working examples |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.