Vercel AI SDK

@pushary/ai-sdk plugs your users' phones into the AI SDK's own tool approval. Before a tool runs, the AI SDK asks an approval function whether it may. Pushary is that function: it sends the tool call to the right user's phone and returns their answer. The tool runs only if they tap Approve.

It works with generateText, streamText and ToolLoopAgent in AI SDK 7.

InstallCopy Anchor Link

npm i @pushary/ai-sdk ai zod
export PUSHARY_API_KEY=pk_xxx.sk_xxx

Get the key at Partner onboarding. The approval gate needs AI SDK 7 and Node.js 22 or later.

Connect a user onceCopy Anchor Link

import { enroll } from '@pushary/ai-sdk'

const { universalLink } = await enroll({ apiKey: process.env.PUSHARY_API_KEY! }, user.id)
// Show universalLink to this user. They open it on their phone once.

user.id is your own id for the person who answers. Your users install the free Pushary app. They never sign up or pay.

Require approval before a tool runsCopy Anchor Link

Pass pusharyApproval() as toolApproval. List the tools that need a yes. Every other tool runs as normal.

import { generateText, stepCountIs, tool } from 'ai'
import { z } from 'zod'
import { pusharyApproval } from '@pushary/ai-sdk'

const issueRefund = tool({
  description: 'Refund an order',
  inputSchema: z.object({ orderId: z.string(), amount: z.number() }),
  execute: async ({ orderId, amount }) => refundOrder(orderId, amount),
})

const lookupOrder = tool({
  description: 'Look up an order',
  inputSchema: z.object({ orderId: z.string() }),
  execute: async ({ orderId }) => findOrder(orderId),
})

const { text } = await generateText({
  model: 'openai/gpt-4o',
  tools: { issueRefund, lookupOrder },
  toolApproval: pusharyApproval({ externalId: user.id, tools: ['issueRefund'] }),
  stopWhen: stepCountIs(5),
  prompt: 'Refund order 1234 in full.',
})

When the model calls issueRefund, the AI SDK stops before execute. Your user gets a notification and taps Approve or Deny. On Approve the refund runs. On Deny, or if nobody answers in time, the tool does not run and the model is told it was not approved. The model cannot skip this, because the AI SDK asks before the tool runs, not the model.

refundOrder and findOrder are your own functions.

Let the model ask a questionCopy Anchor Link

Sometimes the agent needs an answer, not a permission: which order, which address. createPusharyTools gives the model an askHuman tool it can call.

import { generateText, stepCountIs } from 'ai'
import { createPusharyTools } from '@pushary/ai-sdk'

const { text } = await generateText({
  model: 'openai/gpt-4o',
  tools: createPusharyTools({ apiKey: process.env.PUSHARY_API_KEY!, externalId: user.id }),
  stopWhen: stepCountIs(10),
  prompt: 'Ask the customer which order they need help with.',
})

The model chooses when to call askHuman, so use it for questions. Use toolApproval above for anything that spends money, deletes data, or leaves your system.

Good to knowCopy Anchor Link

• Each approval waits up to 55 seconds. For answers that may come later, see callbacks and resume.
• The same tool call never asks twice. A retry reuses the first question.
• Take externalId from your signed-in user, never from the model's tool input.

Source and more examples: @pushary/ai-sdk on GitHub. AI SDK reference: Tool calling and tool approval.