Deploy AI agents with human judgment built in.
The companies deploying AI agents first will lead their sectors. patxin gives your agents access to verified humans on demand — external experts or your own team, via API. Matching in under 60 seconds.
There's a massive opportunity between AI capability and AI deployment.
AI capabilities advance exponentially. Most organizations are still deploying linearly — waiting for compliance frameworks, organizational readiness, enough trust. That gap is where the first movers win.
Your agents already handle 90% of the work. The remaining 10% — the ambiguous clause, the edge-case judgment call, the decision that needs a human signature — that's what separates companies shipping AI workflows from companies still planning them.
Four reasons your competitors haven't deployed yet.
Each one is a barrier for them — and an opportunity for you.
Regulation
Financial services, healthcare, legal — regulators require human oversight of AI decisions. With patxin, every agent decision routes through a verified human with a full audit trail. You deploy while competitors wait for approval.
Organizational readiness
AI at scale means changing workflows, retraining teams, redesigning approval chains. patxin requires zero org change — it works with your team as it is today.
Trust
Decision-makers want a human to verify AI output before it ships. patxin gives your agents that second pair of eyes, on demand, in under 60 seconds.
Integration
Connecting agent output to the right human in real time — not via Slack, not via Jira, not via a contractor marketplace. patxin is one API call.
Three steps. One API call.
Agent calls the patxin API
Your agent hits a decision point. It sends a request with the skill required, urgency, context, and routing preference.
POST /v1/requests
{
"skill": "contract_review_eu",
"urgency": "5min",
"context": "Review termination clause against EU labor directive 2019/1152",
"routing": "internal_first"
} patxin routes to the right human
The routing engine finds the best match based on skill, availability, rating, and your policies. Internal employees are checked first (if configured). External patxers fill the gap. Match time: under 60 seconds.
Human resolves, agent continues
The matched human completes the micro-task (2-30 minutes). The result is delivered back to your agent via webhook or polling. The agent resumes autonomously. No context lost.
One API. Three routing modes.
You choose how tasks get routed.
Marketplace
For teams that need on-demand external expertise.
- Verified external patxers across 20+ skill categories
- Pay per task with credits — no subscriptions
- 15% platform fee on completed tasks
- Patxers are skill-verified and quality-rated
Self-serve. Get started free
In-Corp Starter
Route HITL tasks to your own employees.
- Route tasks to your team via the same API
- Per-seat billing — pay for infrastructure, not your employees' time
- Full policy engine: context controls, PII redaction, skill verification
- Competitor exclusion lists
Custom. Talk to us
In-Corp Scale
High-volume internal routing with usage-based economics.
- Everything in In-Corp Starter
- Low base fee + per-task metering above threshold
- Volume discounts for enterprise contracts
- Dedicated support and custom SLA
Custom. Talk to us
Built for production. Not for demos.
Everything you need to deploy AI agents with human verification from day one.
Context Shield
Define information policies per project: what context the human sees, what gets redacted, what's forbidden. PII auto-redaction, custom regex patterns, URL filtering, code block removal.
Skill verification
Lightweight cognitive assessments. Patxers pass short skill tests before they're matchable. Auto-grading, anti-cheat timing analysis, 80% pass threshold. Valid for 180 days.
Real-time matching
Under 60 seconds from API call to matched human. Multi-candidate notification. Auto-decline timeout with re-matching. Schedule-aware: respects working hours and timezone.
Competitor exclusion
Define which companies your employees cannot accept external tasks from. Post-employment non-compete enforcement. Industry-level exclusions via CNAE/NACE codes.
Routing modes
Three modes per project: external_only (marketplace), internal_first (try your team, fall back to marketplace), internal_only (only your employees).
The dual-role advantage
Employees enrolled via In-Corp can optionally activate as external patxers during non-working hours. Your team gets a side-income channel. patxin's supply side grows organically.
4-tier dispute resolution
AI triage (instant), peer jury (3 verified patxers vote), specialist review, board decision. Automated for tasks under 5 EUR. Transparent and auditable.
Regulatory compliance
Every task creates a documented audit trail: who reviewed it, when, what the decision was, under what policy. GDPR compliant with DPA on registration.
Team management
Five roles: owner, admin, developer, billing, viewer. Role-based permissions for projects, API keys, policies, analytics, and billing. Project-scoped developer access.
Questions companies ask us
Where does my data go when a task is routed to an external patxer?
You control this via information policies. Options range from "open" (patxer sees full context) to "redacted" (PII auto-removed) to "sandboxed" (isolated environment). Context is encrypted in transit and at rest. Sessions are ephemeral.
How do I set up my employees for In-Corp routing?
Add them as team members in the dashboard. Assign roles. Employees receive an invitation, create their account, and become routable for internal tasks. No additional software — tasks arrive via dashboard or Telegram.
How does billing work for In-Corp vs Marketplace?
They're separate. Marketplace uses prepaid credits (pay per task). In-Corp uses subscription billing (per-seat or usage-based). A single company can use both simultaneously.
What happens when an employee leaves?
Their team access is deactivated. Their personal patxer account and reputation persist independently. Post-employment exclusion rules are enforced automatically for the configured duration and then expire.
Can employees opt out of external tasks?
Yes. External marketplace availability is opt-in only. The company enrolls employees for internal routing. Employees separately choose whether to accept external tasks during their own time.
Can I restrict which skills are available per project?
Yes. Each project has an allowed_skills list. Combined with routing mode and information policy, you have fine-grained control over what each agent can do, who it can reach, and what data it can share.
Marketplace (self-serve)
Start routing tasks in 5 minutes. Create an account, buy credits, and make your first API call. No contracts. No sales calls.
Get started freeIn-Corp (enterprise)
Let's design your HITL infrastructure. We'll walk you through routing modes, policy configuration, and pricing.
Talk to us →