# ProspectForge is prospecting intelligence. AI Agent Connect

> ProspectForge replaces static databases with live intelligence. Instead of a list of names and guessed emails, your agent gets a ranked list of prospects backed by an evidence ledger. It pulls from official websites, legal registers, and web archives to confirm roles, company growth signals, and contact details. You get the truth about who is hiring, who is changing their pricing, and who actually holds the roles you need to reach.

## Overview
- **Category:** lead-generation
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_Dpd7CXKRRu0APHa91BTbNtCdGIXvgs05KVsyMcda/ai-agent-connect
- **Tags:** prospecting, lead-generation, outbound, sales-intelligence, b2b, verification, hiring-signals, technographics, keyless, evidence

## Description

You are tired of buying lead lists that are outdated the moment they hit your CRM. ProspectForge changes how you build outbound lists by treating every data point as a claim that requires proof. When you ask for prospects, your agent doesn't just give you a name; it gives you a confidence score based on what it actually found on the live web. If a company is hiring for sales roles or just updated its pricing page, you'll see it. If a contact's role is no longer listed on the company's own team page, the MCP reports it as absent rather than guessing. This is about moving from volume to precision. You can verify a specific company's legal status in a specific jurisdiction, check if a person still holds their title, or even recover old pricing pages from the web archive to see how a company's strategy has shifted. It's a tool built for people who need to know exactly what they can trust before they send a single email.

## Tools

### recover_page
This tool pulls historical pages from Common Crawl indexes. Use it to see old pricing or discontinued products that are no longer on the live site.

### find_people
This tool identifies named individuals in specific roles from company leadership and team pages. Results are heuristic and should be followed up with contact verification.

### find_prospects
This tool discovers candidate companies that match your specific profile. It provides source and fit data without full verification, making it a starting point for deeper research.

### verify_company
This tool generates an evidence ledger for a specific company. It checks the official website, legal entity status, engineering activity, and web archive history.

### verify_contact
This tool confirms if a specific person still holds their role by checking the company's own website. It only reports emails that are actually published on the site.

### verify_legal_entity
This tool checks the GLEIF database to confirm a brand's legal name, jurisdiction, and status. It is useful for verifying a company before signing contracts.

### verify_website
This tool reads official company pages like about, team, and careers to find published emails and social links. It provides the ground truth for contact data.

### company_signals
This tool extracts observable growth signals like open commercial roles, engineering repository activity, or recent changes to pricing pages.

### forge_prospects
This tool is the primary pipeline for building ranked prospect lists. You provide a profile including country, industry, and employee count to get a list of prospects with an evidence ledger for every field.

### similar_prospects
This tool finds companies that resemble a reference brand within a specific region and size band. It uses the reference company to anchor the search when industry data is thin.

## Prompt Examples

**Prompt:** 
```
Find SaaS companies in Portugal with 10-100 employees and Heads of Sales.
```

**Response:** 
```
3 prospects forged. Acme Pay is verified with a confidence of 78/100. It is a fintech company in Portugal with 37 employees. They are currently hiring 4 commercial roles and their pricing page has changed 3 times in 6 months. Inês Costa is confirmed as Head of Sales on the leadership page. Note that no email was published for her on the site.
```

**Prompt:** 
```
Find companies like Stripe in Europe with sales leaders I can contact.
```

**Response:** 
```
5 prospects forged. PayX is the top match with a confidence of 71/100. It is a payments infrastructure company in Portugal with 54 employees. They are hiring 3 commercial roles and a VP Sales is listed on their leadership page. The tool used a keyword fallback because the reference company had limited industry data.
```

**Prompt:** 
```
Does João Silva still work as Head of Sales at this company? Check their site.
```

**Response:** 
```
Contact verification for João Silva: His role as Head of Sales is confirmed on the company's leadership page. No email was found for him on the site. The tool only reports emails that are explicitly published.
```

## Capabilities

### Evidence-based ranking
Your agent ranks prospects based on how much data was actually confirmed versus just guessed.

### Real-time signal detection
The MCP identifies active hiring and website changes to show you companies in growth mode.

### Legal entity verification
Your agent can check official public registers to confirm a company's legal name and jurisdiction.

### Historical web retrieval
The tool accesses web archives to find old product pages or previous contact information.

### Zero-hallucination contact checks
The MCP refuses to invent emails, reporting only what is visibly published on a company's site.

## Use Cases

### Targeting growing companies
Use company signals to find businesses that are actively hiring for commercial roles.

### Verifying contact accuracy
Check if a specific person still holds their position before you start a personalized campaign.

### Competitive intelligence
Use the recover page tool to see how a competitor's pricing or product offerings have changed over time.

### Building lookalike audiences
Find companies that operate like a successful reference brand within your target region.

## Benefits

- Eliminates guesswork by providing an evidence ledger for every field.
- Reduces wasted outreach by verifying roles against official company websites.
- Identifies timing signals like hiring trends and website updates.
- Provides legal certainty through direct checks against public registers.

## How It Works

The MCP follows a structured pipeline to turn raw web data into ranked intelligence.

1. Connect your preferred MCP-compatible client to Vinkius.
2. Describe your ideal customer profile to your agent.
3. The MCP runs a discovery and verification pipeline.
4. Your agent receives a ranked list of prospects with source URLs and confidence scores.

## Frequently Asked Questions

**Do I need an API key or an account?**
No. ProspectForge needs no API key, no account and no credit card. It works entirely from public web data, out of the box.

**Will it give me an email for every contact?**
Only when a public source publishes it. ProspectForge never constructs an email from a name-pattern guess: if nothing publishes one, the field is reported as absent and the answer points at enrichment. That is deliberate — an invented email costs you sender reputation and trust, while a known absence tells you exactly which step to buy.

**How can results be trusted? What does the verification state mean?**
Every field the pipeline asserts lands in an evidence ledger with one of four states: confirmed (at least one independent public source shows it), unverified (a source reported it but no source corroborates it), conflicted (sources disagree — flagged, never averaged away) or absent (no source had it). A whole prospect is verified only when its website is live AND its legal entity is confirmed in the public register; partially_verified means one of the two. Ranking follows confidence: fit (40) + observable signals (35) + verification depth (25).

**What happens when discovery finds few companies for my profile?**
Discovery is strongest for companies with a real public web footprint — a live site, published people pages and visible hiring activity. Many small, purely non-technical businesses leave few public traces, and the result tells you exactly which part of the profile came up thin, so you can widen the employee band, loosen the keywords or try a neighbouring market.

**Is using contact data from company websites allowed?**
Reading publicly published B2B contact data for outreach is generally permitted under GDPR's legitimate-interest basis when there is a real commercial purpose and the individual could reasonably expect it, and you must honor opt-outs. This server reads only public pages and labels every contact with its source and confidence. It is your responsibility to apply that legal basis in your jurisdiction — the tool gives evidence, not legal advice.
