🤖 Built for AI Agents

Give Your AI Agents Reliable Web Access

WebPerception API handles JavaScript rendering, anti-bot bypasses, and structured data extraction — so your agents can focus on reasoning, not scraping.

Get Free API Key → Read Docs

Why agents fail at web tasks

You've built a smart agent. It reasons well. But the moment it needs to read a webpage, everything breaks.

🚫 JavaScript-rendered content

Simple HTTP requests return empty pages. SPAs, React apps, and dynamic sites need a real browser — which your agent doesn't have.

🛡️ Anti-bot protection

Cloudflare, DataDome, reCAPTCHA. Your agent gets blocked on the first request. No amount of prompt engineering fixes this.

🗑️ Messy HTML parsing

Raw HTML wastes tokens and confuses LLMs. Navbars, footers, ads — your agent can't find the actual content in the noise.

⏱️ Infrastructure overhead

Running headless browsers, managing proxies, handling retries — that's a full engineering project, not an agent tool.

One API call. Clean, structured data.

WebPerception gives your agent a single tool that handles everything.

📸 Screenshots & Visual Understanding

Full-page screenshots that your vision-enabled agent can analyze. See the web exactly as a human would.

📄 Clean Text Extraction

Get the main content of any page — no HTML, no nav, no ads. Just the text your agent needs to reason about.

🧠 AI-Powered Data Extraction

Send a prompt describing what you need. Get structured JSON back. Prices, reviews, contact info — whatever the page contains.

🌐 JavaScript Rendering + Anti-Bot Bypass

Every request runs through a real browser with residential proxies. Cloudflare, SPAs, dynamic content — all handled.

Add web access in 3 lines

from langchain.tools import Tool
import requests

def web_scrape(url: str) -> str:
    """Scrape any webpage and return clean text."""
    resp = requests.post(
        "https://api.mantisapi.com/v1/scrape",
        headers={"Authorization": "Bearer YOUR_API_KEY"},
        json={"url": url, "extract": "text"}
    )
    return resp.json()["text"]

# Add to your LangChain agent
web_tool = Tool(
    name="web_scrape",
    func=web_scrape,
    description="Scrape a webpage and return its content as clean text"
)

Works with every agent framework

REST API means universal compatibility. If your framework can make HTTP requests, it works with WebPerception.

🦜

LangChain

Custom tool in 5 minutes. Works with any LangChain agent type.

👥

CrewAI

Give any crew member web research capabilities instantly.

🤖

AutoGen

Multi-agent web tasks with structured data passing.

🔗

OpenAI Functions

Perfect as a function call tool for GPT-4 agents.

Custom Agents

Simple REST API. Python, JavaScript, Go, Rust — any language.

🔄

n8n / Make

No-code agent workflows with HTTP request nodes.

Built for production agents

99.5%
Uptime SLA
<3s
Avg Response
95%+
Success Rate
100
Free Calls/Mo

Ship agents that can actually browse the web

Free tier. No credit card. Start scraping in 30 seconds.

Get Free API Key →