How to Find Micro-Influencers with an API: Automated Instagram Research for AI Agents

Manual influencer research is slow, inconsistent, and impossible to scale. This guide shows how to automate micro-influencer discovery using a pay-per-request Instagram API — searchable by niche, country, follower count, and gender, with no monthly fees or signups.

Why Micro-Influencers Outperform Mega-Influencers

The influencer marketing industry has a signal problem. Accounts with 10M+ followers post sponsored content constantly — their audiences have developed immunity to it. Engagement rates on mega-influencer posts often fall below 1%. Micro-influencers, by contrast, post to tight-knit communities that actually trust their recommendations.

For brands selling niche products — protein supplements, specialty coffee, indie software tools, travel gear — a fitness creator with 25,000 highly engaged followers in the right country is more valuable than a celebrity post that will be skipped by 99% of viewers.

The challenge is finding them at scale.

The Problem with Manual Influencer Research

The typical workflow for finding micro-influencers looks like this:

1. Search a hashtag on Instagram (e.g. #fitnessmotivation)
2. Manually browse 50–100 profiles
3. Check follower counts, engagement rates, content quality
4. Copy usernames into a spreadsheet
5. Search for contact emails on each profile
6. Repeat for every niche and country combination

For a campaign targeting fitness micro-influencers across Germany, France, and the UK, this can take 8–12 hours per country — and you are working with a static snapshot that goes stale as soon as you stop browsing.

Agencies that do this at scale either employ teams of virtual assistants or pay $500–$2,000/month for platforms like Modash or Heepsy. For indie developers and small marketing teams, neither option is practical.

Social Intel API: How It Works

The Social Intel API is a REST endpoint that searches Instagram profiles using real demographic and keyword data. It uses the x402 payment protocol — instead of an API key and a billing dashboard, you include USDC payment in the request header and the API settles on-chain before returning data.

The base endpoint is:

GET https://socialintel.dev/v1/search

On first call, you will receive an HTTP 402 response with payment instructions. Your x402-compatible HTTP client (or the MCP server) handles the payment automatically and retries. The server verifies the on-chain transaction and returns the influencer data.

Available Search Filters

Results are returned sorted by follower count (descending). Each result includes username, full name, follower count, following count, post count, category, biography, public_email (when available), and a direct profile link.

Quickstart: First API Call in 5 Minutes

You need an x402-compatible HTTP client to make paid calls. The easiest option is the official Python SDK:

pip install x402

Then fund a wallet with USDC on Base — any amount over $0.10 is enough to test. The SDK will prompt you to set up a wallet on first run.

import httpx
from x402.httpx import PaymentClientMiddleware

# Configure your wallet private key (Base network, USDC)
client = httpx.Client(
    transport=PaymentClientMiddleware(
        private_key="0x..."  # your wallet private key
    )
)

response = client.get(
    "https://socialintel.dev/v1/search",
    params={
        "query": "fitness",
        "country": "United States",
        "min_followers": 10000,
        "limit": 20
    }
)

for influencer in response.json()["results"]:
    print(influencer["username"], influencer["followers"], influencer.get("public_email"))

Python Example: Find and Export Influencers

This example searches for female fitness micro-influencers in Germany and exports the results — including public emails — to a CSV file:

import httpx
import csv
from x402.httpx import PaymentClientMiddleware

client = httpx.Client(
    transport=PaymentClientMiddleware(private_key="0x...")
)

# Search: female fitness micro-influencers in Germany
resp = client.get(
    "https://socialintel.dev/v1/search",
    params={
        "query": "fitness",
        "country": "Germany",
        "gender": "woman",
        "min_followers": 10000,
        "limit": 100
    }
)
results = resp.json().get("results", [])

# Filter to true micro range: 10K–100K followers
micro = [r for r in results if 10_000 <= r["followers"] <= 100_000]

# Export to CSV
with open("influencers.csv", "w", newline="") as f:
    writer = csv.DictWriter(f, fieldnames=[
        "username", "full_name", "followers",
        "category", "public_email", "biography"
    ])
    writer.writeheader()
    for r in micro:
        writer.writerow({
            "username": r.get("username"),
            "full_name": r.get("full_name"),
            "followers": r.get("followers"),
            "category": r.get("category"),
            "public_email": r.get("public_email", ""),
            "biography": r.get("biography", "")[:100]
        })

print(f"{len(micro)} micro-influencers exported. Total cost: $0.10")

Using the MCP Server with Claude

If you use Claude Desktop or Cursor, you can search for influencers directly in your AI assistant using natural language — no code required. The Social Intel MCP server exposes the search_leads tool, and Claude handles payment automatically from its agent wallet.

Add the MCP server to your configuration:

{
  "mcpServers": {
    "social-intel": {
      "command": "python",
      "args": ["-m", "mcp_server"],
      "env": {
        "SOCIAL_INTEL_API_URL": "https://socialintel.dev",
        "WALLET_PRIVATE_KEY": "0x..."
      }
    }
  }
}

Then in Claude, you can simply ask:

Claude will call search_leads, pay $0.10 USDC automatically, and return the structured results in the chat window. You can then ask it to filter, sort, or format the data however you need.

Getting Emails for Direct Outreach

Instagram business accounts often have a public contact email on their profile — the one they use for brand partnerships. The Social Intel API returns this as public_email in the response when it is present.

In our testing with fitness business accounts:

• ~30–35% of business accounts have a public email
• Personal accounts rarely have public emails
• Verified accounts are more likely to have contact info
• Emails are formatted as-is from the Instagram profile — no scraping or inference

For a list of 100 results, you can expect roughly 30–35 usable direct contact emails without any additional enrichment. Combined with the username, follower count, and category data, this gives you a working outreach list in a single API call.

Pricing Breakdown

The API uses dynamic per-request pricing via the x402 protocol:

There are no monthly minimums, no rate limits, and no signup required. If you make 10 calls per month searching different niches and countries, you spend approximately $1.00–$2.60 total.

Payment requires USDC on Base chain. The x402 client handles payment automatically — you load a wallet once and it settles on-chain per request. See the x402 guide for wallet setup details.

Frequently Asked Questions

What is a micro-influencer?

Micro-influencers are creators with 10,000–100,000 followers. They typically have higher engagement rates (3–8%) than mega-influencers, more niche audiences, and accept collaborations at lower cost. For brands targeting specific demographics, micro-influencers often deliver better ROI than celebrity partnerships.

How do I search Instagram influencers by niche with an API?

The Social Intel API accepts a free-text query parameter (e.g. fitness, vegan cooking, travel photography) and returns Instagram accounts ranked by follower count. You can combine it with country, gender, min_followers, and city filters to narrow results. A single API call returns up to 100 profiles with username, follower count, category, and public email for business accounts.

Does the Social Intel API return email addresses?

Yes — for Instagram business accounts that have a public contact email visible on their profile, the API returns the public_email field. Personal accounts typically do not have public emails. In our tests, roughly 1 in 3 fitness business accounts had a public email address available.

How much does the micro-influencer API cost?

Pricing is $0.10 USDC for up to 20 results, or $0.10 + $0.002 per result above 20 (max 100 results). There are no monthly fees, no signup, and no minimum spend. You pay only when you make a request.

Can I use this API with Claude or Cursor to find influencers automatically?

Yes. The Social Intel MCP server lets Claude Desktop, Cursor, and other MCP-compatible AI assistants search for Instagram influencers directly. The AI agent handles payment automatically — you just ask in natural language: "Find 20 fitness micro-influencers in Germany with at least 15,000 followers."

What filters are available for influencer search?

Available filters: query (keyword), country (full name), gender (man or woman), min_followers, city, is_verified, and limit (1–100). Results are sorted by follower count descending.

Related: What is x402? How AI agents pay for APIs · Using Instagram data in Claude via MCP · Instagram Data APIs in 2026 Compared