When your agents hit human-only gates, SiliconBridge handles them via API.
Your agent keeps running.
Partnership program for Browserbase, Skyvern, Airtop, and other AI agent platforms.
Your agents are powerful. Until they hit a wall.
reCAPTCHA v3, hCaptcha, image challenges — agents fail without human eyes.
SMS codes, call verification, authenticator apps — only humans can receive them.
Contracts, agreements, legally binding documents — need real humans.
KYC, background checks, biometric scans — human judgment required.
Opening bank accounts, crypto wallets, payment providers — compliance gates require humans.
Outbound calls, customer support, verbal authorization — no way for agents to handle this alone.
Today: Agents get stuck. Workflows fail. You need manual intervention or workarounds.
With SiliconBridge: Agents call a human, wait 30 seconds, and keep running.
3 lines of code. That's all.
import httpx
result = httpx.post("https://siliconbridge.xyz/task/submit",
headers={"X-API-Key": api_key},
json={"service_type": "captcha_solve", "payload": {"image_url": captcha_url}}
).json()
# Human solves it in ~30 seconds, agent continues
agent.proceed_with(result["solution"])Call SiliconBridge when your Browserbase session hits a bot check or manual gate.
from browserbase import Browserbase
import httpx
bb = Browserbase()
sb_key = "sk_live_xxx"
try:
result = bb.get("https://example.com")
except BotCheckError:
# Agent hit a gate
captcha = bb.screenshot_region(...)
# Call SiliconBridge
solved = httpx.post(
"https://siliconbridge.xyz/task",
headers={"X-API-Key": sb_key},
json={
"service_type": "captcha_solve",
"payload": {"image_url": captcha}
}
).json()
# Continue
result = bb.type(solved["solution"])Reliable browser sessions with human fallback. Perfect pair.
Your browser automation gets human judgment. No restarts, no manual retries.
AI-driven browser automation meets human operators. Unstoppable.
Skyvern's intelligent navigation + SiliconBridge's human gates = complete autonomy.
Add SiliconBridge as a fallback task type in your Skyvern agent.
from skyvern import Agent
import httpx
agent = Agent()
sb_key = "sk_live_xxx"
# When Skyvern's LLM hits a dead end
@agent.on_task_blocked
def handle_human_gate(context):
# Describe to SiliconBridge
task = httpx.post(
"https://siliconbridge.xyz/task",
headers={"X-API-Key": sb_key},
json={
"service_type": "human_approval",
"description": context.task_description,
"screenshot": context.screenshot_url,
"timeout": 60
}
).json()
return task["result"]
agent.run("https://example.com/workflow")Use SiliconBridge as a tool in your Airtop multi-agent system.
from airtop import Agent, Tool
import httpx
class SiliconBridgeTool(Tool):
name = "siliconbridge_human"
description = "Call human operators for gates"
def execute(self, task_type, payload):
result = httpx.post(
"https://siliconbridge.xyz/task",
headers={"X-API-Key": os.getenv("SB_KEY")},
json={
"service_type": task_type,
"payload": payload
}
).json()
return result
agent = Agent(tools=[SiliconBridgeTool()])
agent.run_on("https://complex-app.com")Multi-agent orchestration meets human-in-the-loop. Seamless.
Your agents coordinate. When they need humans, SiliconBridge slots in as a tool.
22 services across 4 categories. All via API.
We complement your platform. We don't compete.
We're not a browser automation tool. We're the human fallback for when bots hit walls. Your platform stays at the center.
Real humans. Real results. When your agent calls, a human picks up and solves the problem in seconds.
No waiting in queue. No delays. Humans are standing by. Median response time: 30 seconds.
Model Context Protocol integration. Add via URL. Works with Claude, any LLM-powered agent, and your custom platforms.
Use what you need. No monthly fees. No commitments. Pricing starts at $1 per task. Volume discounts available.
Pay in USD, EUR, or crypto. L402 machine-native billing for on-chain agents. Full automation.
Pick your path: MCP, REST API, Python SDK, JS SDK, or framework tools (LangChain, CrewAI).
Dedicated partnership contact. Co-marketing. Priority support. Revenue sharing opportunities.
Pick the path that fits your architecture.
Plug into Claude, any LLM, or your custom agent.
"mcp": {
"siliconbridge": {
"command": "uvx",
"args": ["siliconbridge-mcp"]
}
}Simple HTTP POST. Works everywhere.
POST /task/submit
X-API-Key: sk_live_xxx
{
"service_type": "captcha_solve",
"payload": {...}
}Drop-in for Python agents and frameworks.
pip install siliconbridge
from siliconbridge import SiliconBridge
sb = SiliconBridge(api_key="...")
result = sb.submit_task("otp", {...})For Node.js, browser, and TypeScript.
npm install siliconbridge
import { SiliconBridge } from 'siliconbridge';
const sb = new SiliconBridge({apiKey: '...'});
const result = await sb.submitTask('otp', {...});Pre-built integrations for popular AI agent frameworks. Use SiliconBridge as a native tool in your agent stack.
Copy, paste, customize.
from browserbase import Browserbase
from siliconbridge import SiliconBridge
bb = Browserbase()
sb = SiliconBridge(api_key="sk_live_xxx")
session = bb.browser_context()
# Your agent workflow
response = session.goto("https://example.com")
# Hit a CAPTCHA? Call SiliconBridge
if "captcha" in response.text.lower():
screenshot = session.screenshot()
result = sb.submit_task("captcha_solve", {
"image_url": upload_to_cdn(screenshot)
})
session.type(result["solution"])
session.click("button[type='submit']")from skyvern.agents import Agent
from siliconbridge import SiliconBridge
sb = SiliconBridge(api_key="sk_live_xxx")
agent = Agent(
instructions="""
1. Navigate to https://apply.example.com
2. Fill the form
3. If you see a CAPTCHA or verification:
- Call tool: siliconbridge_human_gate
- Pass the page context
- Wait for result, then proceed
"""
)
# Register SiliconBridge as a tool
agent.register_tool(
name="siliconbridge_human_gate",
description="Solve human-only gates (CAPTCHA, 2FA, phone calls)",
handler=lambda gate_type, context: sb.submit_task(
gate_type,
context
)
)
result = agent.run()from airtop import Agents
import httpx
SB_API_KEY = "sk_live_xxx"
agents = Agents(
agents=[
{
"name": "form_filler",
"role": "Fill out the registration form"
},
{
"name": "verification_handler",
"role": "Handle CAPTCHA, OTP, phone verification"
}
]
)
# When verification_handler hits a gate
@agents.tool("siliconbridge")
def call_human_operator(gate_type: str, payload: dict):
"""Call human operator for verification gates"""
response = httpx.post(
"https://siliconbridge.xyz/task/submit",
headers={"X-API-Key": SB_API_KEY},
json={"service_type": gate_type, "payload": payload}
)
return response.json()["result"]
# Run your multi-agent workflow
agents.run("https://complex-onboarding.example.com")Browserbase. Skyvern. Airtop. Any agent platform. SiliconBridge handles the human gates. Your agents never get stuck.
Partnership inquiries, integration questions, or custom arrangements.
Get in Touchxsiliconbridgex@gmail.com