> ## Documentation Index
> Fetch the complete documentation index at: https://developers.t2000.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Prep for a meeting

> A name and a company in, a tight pre-meeting brief out. Perplexity + Exa + Serper pull the company, the person, and what's recent — Claude flags what matters. Paid as you go, no API keys.

You're meeting someone in an hour and want the 90-second version: what the company is, who the person is, and anything recent worth knowing. Your agent pulls a web-grounded company brief, a profile of the person, and the latest news, then Claude distills it into a brief and flags the relevant bits. \~\$0.08, self-contained — just a name, a company, and the topic.

<Note>
  This is a **durable gateway demo** — a sandboxed client can't fake live, web-grounded research. The value is paid search + real-time data, not generation. (We have no X/Twitter endpoint today, so "recent posts" reads as recent public posts + news; an X route would sharpen this — see the recipe backlog.)
</Note>

***

## The prompt

```
Use t2 services. I'm meeting Jensen Huang from NVIDIA about AI infrastructure. Use Perplexity
for a company brief (size, funding/financials), Exa for a profile on him (role,
background), and Serper for recent news — then flag anything relevant. Keep it
to one page.
```

***

## What runs

1. `POST /perplexity/v1/chat/completions` — web-grounded company brief with citations (\~\$0.02)
2. `POST /exa/v1/search` — semantic search for the person's profile + recent public activity (\~\$0.02)
3. `POST /serper/v1/search` — latest news mentioning both (\~\$0.02)
4. `POST /anthropic/v1/messages` — Claude synthesizes the one-pager and flags what's relevant (\~\$0.02)

***

## Run it

### Claude Desktop (MCP)

```bash theme={null}
npm install -g @t2000/cli && t2 init && t2 fund && t2 mcp install
```

Paste the prompt with any person + company. The agent decides how many source passes it needs.

### SDK

```typescript theme={null}
import { T2000 } from '@t2000/sdk';

const agent = await T2000.create();
const person = 'Jensen Huang';
const company = 'NVIDIA';
const topic = 'AI infrastructure';

const [companyBrief, profile, news] = await Promise.all([
  agent.pay({
    url: 'https://mpp.t2000.ai/perplexity/v1/chat/completions',
    method: 'POST',
    body: JSON.stringify({
      model: 'sonar',
      messages: [{
        role: 'user',
        content: `Brief on ${company}: what it does, size, funding/financials, recent direction. Cite sources.`,
      }],
    }),
  }),
  agent.pay({
    url: 'https://mpp.t2000.ai/exa/v1/search',
    method: 'POST',
    body: JSON.stringify({
      query: `${person} ${company} role background recent`,
      numResults: 5,
      contents: { text: true },
    }),
  }),
  agent.pay({
    url: 'https://mpp.t2000.ai/serper/v1/search',
    method: 'POST',
    body: JSON.stringify({ q: `${person} ${company} news` }),
  }),
]);

const brief = await agent.pay({
  url: 'https://mpp.t2000.ai/anthropic/v1/messages',
  method: 'POST',
  headers: { 'anthropic-version': '2023-06-01' },
  body: JSON.stringify({
    model: 'claude-sonnet-4-5',
    max_tokens: 800,
    messages: [{
      role: 'user',
      content:
        `One-page pre-meeting brief. I'm meeting ${person} from ${company} about ${topic}. ` +
        `Sections: Company (1 para), Person (1 para), Recent & relevant (3-5 bullets), ` +
        `2-3 smart questions to ask. Keep every claim tied to a source.\n\n` +
        `COMPANY: ${JSON.stringify(companyBrief.body)}\n\nPROFILE: ${JSON.stringify(profile.body)}\n\nNEWS: ${JSON.stringify(news.body)}`,
    }],
  }),
});

console.log((brief.body as { content: { text: string }[] }).content[0].text);
```

***

## Expected output

```
4 calls · ~$0.08 · ~9s · 0 taps
One-page brief: company + person + recent + questions to ask
```

***

## Extend it

* Swap **Serper** for **Brave news** (`/brave/v1/news/search`) or **NewsAPI** (`/newsapi/v1/search`) for a different recency lens
* Add **Hunter.io** (`/hunter/v1/search`) to surface a verified contact email for follow-up
* Pipe the brief into **Resend** (`/resend/v1/emails`) to send yourself the prep the morning of
* Run it for every attendee on tomorrow's calendar in parallel — each is an independent \~\$0.08 pass