mcpField
D
deepagents-typescript-quickstart

deepagents-typescript-quickstart

github.com/langchain-aiAgent workflows1.3k5.1k installsUpdated 3d ago
Repository

About deepagents-typescript-quickstart

A skill for implementing the Deep Agents TypeScript research-agent quickstart. It supports provider selection and uses provider-provided web search rather than Tavily.

Key features

  1. 01Official quickstart-based implementation
  2. 02Provider and model selection
  3. 03Provider built-in web search
  4. 04Creates work in a new directory
  5. 05Runs the research-agent example

Use cases

  1. 01Starting a Deep Agents TypeScript project
  2. 02Running a research-agent example without Tavily

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/langchain-ai/langchain-skills --skill deepagents-typescript-quickstart

After installing, type/deepagents-typescript-quickstartin Claude Code to confirm it loaded.

Details

From the project README

Deep Agents TypeScript quickstart

Follow the live docs — do not invent an alternate API from memory:

https://docs.langchain.com/oss/javascript/deepagents/quickstart

Fetch that page (Docs MCP or HTTP) and implement the research-agent shape it shows (createDeepAgent, research system prompt, invoke with a research question like “What is LangGraph?”). Requires Node 22+.

Local setup constraints

Apply these on top of the quickstart (they keep setup minimal and model-agnostic):

  1. Ask which provider/model to use. Showcase that Deep Agents are model-agnostic. Suggested prompt:

    Which model should this agent use? Pass a provider:model string — e.g. openai:gpt-5.5, anthropic:claude-sonnet-5, google-genai:gemini-3.5-flash. Default if you're unsure: anthropic:claude-sonnet-5.
    We'll use that provider's built-in web search (no separate search API key).

  2. Create a new directory (e.g. deep-agent/) and do all work there — do not pollute the open project.

  3. Do not use Tavily (or @langchain/tavily). Replace the quickstart's search tool with the chosen provider's built-in web search. Look up the current export/tool shape on that provider's LangChain docs (examples as of writing — re-check if needed):

    ProviderBuilt-in search tool
    Anthropic@langchain/anthropic tools.webSearch_*() (or equivalent dict)
    OpenAI{ type: "web_search" }
    Google{ google_search: {} }

    Prefer Anthropic / OpenAI / Google so provider search is available. Only secret: that provider's API key in .env (gitignored). Skip LangSmith tracing unless they ask.

  4. Install packages from the quickstart minus Tavily; add the provider package for their model.

  5. Run the research example, show output, then stop. Point to deep-agents-core / customization / Managed Deep Agents for next steps.

Comments

Markdown supported