Vexi — Structured Business Data Built for Agents

02 · Live request

See it in action.

One endpoint. Predictable shape. Cached, confidence-scored, agent-ready.

request.shGET

curl -X GET "https://api.getvexi.dev/v1/search" \
  -H "Authorization: Bearer $VEXI_KEY" \
  -d '{
    "query": "scheduling software",
    "location": "San Francisco",
    "limit": 1
  }'

response.json 187 ms · cache hit

{
  "results": [{
    "name": "Cal.com",
    "business_type": "saas",
    "description_short": "Open-source scheduling infrastructure for everyone.",
    "offerings": {
      "services": [
        { "name": "Team Scheduling", "category": "Scheduling" },
        { "name": "Enterprise SSO",  "category": "Security"   }
      ],
      "pricing": {
        "model": "freemium",
        "starting_from": "$12/month",
        "free_tier": true
      }
    },
    "contact": {
      "website":  "https://cal.com",
      "email":    "support@cal.com",
      "linkedin": "https://linkedin.com/company/calcom"
    },
    "trust": {
      "differentiators": [
        "Open source with 30k+ GitHub stars",
        "Self-hostable — full data control"
      ],
      "ratings": [{ "platform": "G2", "score": "4.7/5" }]
    },
    "agent_interface": {
      "actions": [
        { "name": "Book a demo",     "url": "https://cal.com/sales",        "method": "link" },
        { "name": "Start free trial", "url": "https://app.cal.com/signup",   "method": "link" }
      ],
      "intents_matched": [
        "schedule a meeting", "book a demo", "find scheduling software"
      ],
      "tags": ["scheduling", "open-source", "saas"]
    },
    "quality": {
      "completeness_score": 94,
      "confidence":         "high",
      "data_freshness":     "fresh"
    }
  }],
  "total_ready": 1,
  "source": "intelligent_search",
  "meta": { "wait_time_ms": 187 }
}

↑ Truncated response. Full ABO includes 40+ structured fields across 8 sections.

Want to see the full ABO?Get your API Key for free

03 · The difference

Every other tool gives your agent raw materials.
Vexi gives it answers.

Search APIs find URLs. Crawlers clean HTML. LLMs extract text. None of them give your agent structured, cached, confidence-scored business data.

Without Vexi

With Vexi

Inconsistency

Every website is different. You wrote custom parsing logic for one crawler, then another, then your LLM started hallucinating fields. A site redesign broke everything on a Friday night.

One schema. Every business. Forever. Restaurant or SaaS, local or global — every ABO returns the same 8-section structure. Write your agent logic once.

Reliability

Your agent confidently recommended a business with a disconnected phone number. Search APIs don't know it changed. LLMs don't know. Your user found out the hard way.

Your agent knows what it doesn't know. Every ABO includes completeness_score, confidence, data_freshness and missing_fields. No more silent failures.

Cost

Crawlers return 40,000 tokens of HTML noise. Search APIs return snippets. LLMs still need to extract, structure, and verify. Every query. Every business. Every time.

200ms. Not 90 seconds. ~$0. Not $0.05.First query generates the ABO. Every query after returns from cache. 95% less tokens. 99.5% less compute.

This is what business data built for agents looks like.

See the full ABO schema →

04 · Flow

From query to Agent Business Object
in three steps.

01

Describe what you need

Search by intent, category, or location in natural language. No URLs required. Your agent asks — Vexi finds.

search_businesses({
  query: "CRM for startups",
  location: "Brazil"
})

02

Vexi resolves and returns

If the business is in the cache, you get a complete ABO in 200ms. If not, Vexi crawls it automatically and returns when ready.

← cache hit: 187ms
← crawling: ~30s

03

Your agent reasons and acts

The ABO contains everything your agent needs — structured data, confidence scores, and executable actions. No parsing. No guessing. No post-processing.

actions[0].url → "https://cal.com/sales"

Easy to set up. Easy to maintain.

No pipelines to build. No schemas to maintain. No parsers to fix.

05 · Built for production

More reasons your agents will love Vexi.

Intent-based discovery

Query by intent, not by URL. Your agent describes what it needs — Vexi finds the right businesses.

Executable actions included

Every ABO includes ready-to-use action URLs. Your agent doesn't just find businesses — it knows exactly what to do next.

A directory that grows with demand

Every query enriches the database for everyone. The more agents use Vexi, the better it gets — automatically.

Native MCP server

Connect once. Your agent discovers and uses Vexi tools automatically — no integration code required.

On-demand enrichment

Tell Vexi which fields your agent needs. If they're missing, Vexi fetches them before returning. Your agent always gets what it asked for.

Data that stays fresh

Stale ABOs are automatically re-crawled. No cache invalidation to manage. No outdated data silently breaking your agent.

Works with every major agent framework.

REST API or native MCP — your choice.

06 · Pricing

Free to start. Priced for production.

No seat fees. No feature gating. You pay for what you use.

Free$0/mo

For developers exploring Vexi

500 credits / month

Full access to the API, MCP server, and complete ABO schema. Everything you need to build and test your agent.

Get started free

No credit card required

The agentic web is being
built right now.

The businesses that get structured first will be the ones agents recommend. Start with 500 free credits — no credit card, no setup, no pipelines to build.

Get your free API key

Takes 30 seconds. Works with LangChain, CrewAI, Claude, and any MCP-compatible agent.