LangChain.js Integration

JACS provides two integration patterns for LangChain.js:

Full toolkit -- expose all JACS operations as LangChain tools your agent can call
Auto-signing wrappers -- transparently sign existing tool outputs

5-Minute Quickstart

1. Install

npm install @hai.ai/jacs @langchain/core

2. Create a JACS client

import { JacsClient } from '@hai.ai/jacs/client';

const client = await JacsClient.quickstart();

3. Sign tool outputs

import { signedTool } from '@hai.ai/jacs/langchain';

const signed = signedTool(mySearchTool, { client });
const result = await signed.invoke({ query: 'hello' }); // result is JACS-signed

Or give your agent the full JACS toolkit:

import { createJacsTools } from '@hai.ai/jacs/langchain';

const jacsTools = createJacsTools({ client });
const allTools = [...myTools, ...jacsTools];
const llmWithTools = model.bindTools(allTools);

createJacsTools() returns an array of LangChain DynamicStructuredTool instances wrapping the full JacsClient API. Bind these to your LLM so the agent can sign, verify, create agreements, manage trust, and audit -- all as part of its reasoning.

import { JacsClient } from '@hai.ai/jacs/client';
import { createJacsTools } from '@hai.ai/jacs/langchain';

const client = await JacsClient.quickstart();
const jacsTools = createJacsTools({ client });

// Combine with your own tools and bind to an LLM
const allTools = [...myTools, ...jacsTools];
const llmWithTools = model.bindTools(allTools);

Available Tools

Tool	Description
`jacs_sign`	Sign arbitrary JSON data with cryptographic provenance
`jacs_verify`	Verify a signed document
`jacs_create_agreement`	Create a multi-party agreement
`jacs_sign_agreement`	Sign an existing agreement
`jacs_check_agreement`	Check agreement status (signatures, completeness)
`jacs_verify_self`	Verify this agent's integrity
`jacs_trust_agent`	Add an agent to the local trust store
`jacs_list_trusted`	List all trusted agent IDs
`jacs_is_trusted`	Check if a specific agent is trusted
`jacs_audit`	Run a security audit
`jacs_agent_info`	Get agent ID, name, and status

Strict Mode

Pass strict: true to make tools throw on errors instead of returning error JSON:

const tools = createJacsTools({ client, strict: true });

Auto-Signing Wrappers

signedTool

Wraps any LangChain BaseTool so its output is automatically signed:

import { signedTool } from '@hai.ai/jacs/langchain';

const signed = signedTool(mySearchTool, { client });
const result = await signed.invoke({ query: 'hello' }); // result is JACS-signed

jacsToolNode (LangGraph)

Creates a LangGraph ToolNode where every tool's output is signed:

import { jacsToolNode } from '@hai.ai/jacs/langchain';

const node = jacsToolNode([tool1, tool2], { client });

Requires @langchain/langgraph.

jacsWrapToolCall

Returns an async wrapper for manual tool execution in custom LangGraph workflows:

import { jacsWrapToolCall } from '@hai.ai/jacs/langchain';

const wrapFn = jacsWrapToolCall({ client });
// Use in custom graph: const result = await wrapFn(toolCall, runnable);

Installation

npm install @hai.ai/jacs @langchain/core
# Optional for jacsToolNode:
npm install @langchain/langgraph

All @langchain/* imports are lazy -- the module can be imported without LangChain installed.

Next Steps

MCP Integration -- Full JACS tool suite for MCP servers
Vercel AI SDK -- AI model provenance signing
Express Middleware -- HTTP API signing

JACS Usage Documentation