Skip to main content
Imagine a virtual sous-chef who listens to your ingredient list and responds with suggestions, techniques, or even a full recipe—all served via a chat UI.

​
What You’ll Build

  • A Mastra agent that thinks like a seasoned home cook.
  • Tools for helpful interactive features—substitute ingredients, create recipes.
  • A conversational UI (React example provided).
  • A deployable agent available via Mastra’s API.

​
Prerequisites

  • A Mastra project (npx create-mastra@latest my-mastra-app).
  • Node.js installed.
  • OpenAI API key in .env as OPENAI_API_KEY.
  • Optional: Familiarity with Zod schema for structured output.

​
Step 1

​
Create the Tools

1

Substitute tool

Create a tool to suggest ingredient substitutions.
2

Recipe tool

Create a tool that generates a simple recipe from a pantry list.
src/tools/suggest-substitute-tool.ts: 
import { createTool } from '@mastra/core/tools';
import { z } from 'zod';

export const suggestSubstituteTool = createTool({
  id: 'suggest-substitute',
  description: 'Suggest a substitute for an ingredient.',
  inputSchema: z.object({
    ingredient: z.string().describe('Ingredient to substitute'),
  }),
  outputSchema: z.object({
    substitute: z.string(),
    reason: z.string(),
  }),
  execute: async ({ context }) => {
    const ingredient = context.ingredient.toLowerCase();
    const map: Record<string, { substitute: string; reason: string }> = {
      milk: { substitute: 'almond milk', reason: 'Similar texture and flavor' },
      butter: { substitute: 'margarine', reason: 'Same fat profile for cooking/baking' },
    };
    return map[ingredient] || {
      substitute: 'No direct match',
      reason: 'Choose an ingredient with similar flavor or texture.',
    };
  },
});
src/tools/recipe-from-pantry-tool.ts: 
import { createTool } from '@mastra/core/tools';
import { z } from 'zod';

export const recipeFromPantryTool = createTool({
  id: 'recipe-from-pantry',
  description: 'Generate a simple recipe using available ingredients.',
  inputSchema: z.object({
    ingredients: z.array(z.string()).min(1),
  }),
  outputSchema: z.object({
    title: z.string(),
    steps: z.array(z.string()),
  }),
  execute: async ({ context }) => {
    const { ingredients } = context;
    return {
      title: 'Pantry Special',
      steps: [
        `Use ${ingredients.join(', ')}.`,
        'Cook and enjoy!',
      ],
    };
  },
});

​
Step 2

​
Create the Agent

src/agents/chef-agent.ts: 
import { openai } from '@ai-sdk/openai';
import { Agent } from '@mastra/core/agent';
import { Memory } from '@mastra/memory';
import { LibSQLStore } from '@mastra/libsql';

import { suggestSubstituteTool } from '../tools/suggest-substitute-tool';
import { recipeFromPantryTool } from '../tools/recipe-from-pantry-tool';

export const chefAgent = new Agent({
  name: 'Chef Agent',
  instructions: `
You are a friendly home chef. Help users cook based on their available ingredients.
Use 'suggest-substitute' to offer alternatives.
Use 'recipe-from-pantry' to build recipes from their list.
Keep answers short and clear.
  `,
  model: openai('gpt-5'),
  tools: {
    'suggest-substitute': suggestSubstituteTool,
    'recipe-from-pantry': recipeFromPantryTool,
  },
  memory: new Memory({
    storage: new LibSQLStore({ url: 'file:../mastra.db' }),
  }),
});

​
Step 3

​
Register the Agent in Mastra

src/mastra/index.ts: 
import { Mastra } from '@mastra/core/mastra';
import { PinoLogger } from '@mastra/loggers';
import { LibSQLStore } from '@mastra/libsql';

import { chefAgent } from '../agents/chef-agent';

export const mastra = new Mastra({
  agents: { chef: chefAgent }, // "chef" is the REST id
  storage: new LibSQLStore({ url: 'file:../mastra.db' }),
  logger: new PinoLogger({ name: 'Mastra', level: 'info' }),
});
The key in agents: { chef: chefAgent } determines the API path: /api/agents/chef/*.

​
Step 4

​
Run the Agent

From your project root: 
rm -rf .mastra/output
npx mastra dev
You should see: 
Mastra API running on port http://localhost:4111/api

​
Step 5

​
Deploy & Production

curl -X POST http://localhost:4111/api/agents/chef/generate \
  -H "Content-Type: application/json" \
  -d '{"messages":[{"role":"user","content":"ping"}]}'

Use a tunnel to test the agent before full deployment:

ngrok http 4111
cloudflared tunnel --url http://localhost:4111
loca.lt --port 4111

Append /api/agents/chef/generate and copy the HTTPS URL into the Dashboard.

// api/agents/chef/generate.ts
import type { VercelRequest, VercelResponse } from '@vercel/node';
import { chefAgent } from '../../../src/agents/chef-agent';

export default async function handler(req: VercelRequest, res: VercelResponse) {
  if (req.method !== 'POST') return res.status(405).json({ error: 'Method not allowed' });
  try {
    const { messages } = req.body || {};
    const result = await chefAgent.respond({ messages });
    return res.status(200).json(result);
  } catch (e: any) {
    return res.status(500).json({ error: e.message || 'Agent error' });
  }
}
  • Serverless: Single lightweight function calling the agent.
  • Container: Long‑lived process; add health checks & logging.
  • Edge: Keep tools stateless; externalize persistence.
  • Rate‑limit (per IP + user).
  • Add auth (Bearer/JWT) for non-public usage.
  • Log tool calls (name, duration, success/error).
Deployment URL = public HTTPS endpoint + /api/agents/chef/generate. Mastra Agent ID = chef.

​
Step 6

​
Configure in CometChat

1

Open Dashboard

Open the CometChat Dashboard.
2

Navigate

Go to your App → AI Agents.
3

Add agent

Set Provider=Mastra, Agent ID=chef, Deployment URL=your public generate endpoint.
4

(Optional) Enhancements

Add greeting, prompts, and configure actions/tools if you use frontend tools.
5

Enable

Save and ensure the agent toggle shows Enabled.

​
Step 7

​
Customize in UI Kit Builder

1

Open variant

From AI Agents click the variant (or Get Started) to enter UI Kit Builder.
2

Customize & Deploy

Select Customize and Deploy.
3

Adjust settings

Theme, layout, features; ensure the Mastra agent is attached.
4

Preview

Use live preview to validate responses & any tool triggers.

​
Step 8

​
Integrate

Once your agent is configured, you can integrate it into your app using the CometChat Widget Builder
Widget

Widget Builder

React

React UI Kit

Pre Built UI Components
Note: The Mastra agent you connected in earlier steps is already part of the exported configuration, so your end-users will chat with that agent immediately.

​
Step 9

​
Test Your Setup

1

API generates response

POST to /api/agents/chef/generate returns a recipe or suggestion.
2

Agent listed

/api/agents includes “chef”.
3

Tool action works

UI handles suggest-substitute tool invocation.
4

Full sample test (run curl)

See curl command below.
curl -X POST http://localhost:4111/api/agents/chef/generate \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      { "role": "user", "content": "I have pasta, tomatoes, garlic — make me dinner" }
    ]
  }'

​
Troubleshooting

  • Agent not found: Check the agents map in mastra/index.ts.
  • Tool not firing: Ensure the id matches exactly between the tool file and the agent.
  • API key errors: Confirm OPENAI_API_KEY is set.

​
Next Steps

  • Add more tools (e.g., nutritional info, shopping list builder).
  • Add a workflow to decide automatically between quick tips and full recipes.
  • Localize outputs for different languages.