Now live — Policy Engine · Memory v2 · Flows · Graph

Infrastructure for autonomous
AI agents.

Enable agents to discover capabilities, access real-world intelligence, and transact automatically — through one unified network.

One integration. Continuous execution. Pay-per-use.

Start building free → View documentation

No credit card required · Works with any agent framework · Setup in ~2 minutes

yantrix — research agent · live
$
Policy Engine · Semantic Memory v2 · Flow Pipelines · Orchestration Graph · x402 USDC on Base · MCP · A2A · Relay · 40 production APIs
// the problem

Agents can reason.
The infrastructure hasn't kept up.

Autonomous agents are powerful. But without governance, they are unpredictable, unauditable, and unsafe in production.

Agents overspend budgets. Forget context between sessions. Use tools they shouldn't. And leave no audit trail when something goes wrong.

Yantrix is the governance and economic control plane that makes agents safe, auditable, and autonomous in production.

Without Yantrix
Agents overspend with no guardrails
No audit trail of agent decisions
Memory lost between sessions
Billing requires human approval
Workflows locked to one framework
With Yantrix
Policy Engine governs every action
Immutable audit trail per agent
Semantic memory persists across sessions
x402 autonomous payments, no approvals
Flows + Graph — framework agnostic
// start in 60 seconds

One command.
Entire network.

Install the MCP package and get 71+ tools in any agent framework. Or call APIs directly with a single header.

1
Register with your wallet — $1 USDC on Base
2
Get your API key — works across all 40 APIs
3
Call any API — first 6 months free

No billing setup required · Works with Claude, GPT-4, any framework

# Install MCP package
npx @yantrixai/mcp
# Call any API directly
curl https://gateway.yantrix.ai/proxy/ifsc/ifsc/HDFC0000001 \
-H "X-API-Key: ytx_..."
{"bank": "HDFC Bank", "cost": "$0.002",
"covered_by_offer": true}
# Discover all capabilities
curl https://gateway.yantrix.ai/v1/catalog
// how it works

Four layers.
Complete agent infrastructure.

Every layer creates a different kind of value. Together they form infrastructure that compounds over time.

Enterprise lock-in

Govern

Policy Engine evaluates every action before execution. Spend limits. Tool permissions. Audit trail. CFOs trust it.

POST /v1/policy/evaluate
Data lock-in

Remember

Semantic Memory v2. pgvector similarity search. Episodic + semantic memory types. Data stays in Yantrix.

POST /memory/search
Operational lock-in

Pipeline

Flow Pipelines with conditions and context passing. Define once. Run autonomously. Agents stop needing LangChain.

POST /v1/flows/create
Developer lock-in

Orchestrate

Stateful multi-agent graphs. Parallel execution. Dependency-aware. Workflows live inside Yantrix — not in agent code.

POST /v1/graphs/create

Works with Claude, GPT-4o, Gemini, any LLM framework

// built-in primitives

Seven capabilities.
Infinite combinations.

Composable building blocks for any autonomous workflow.

Identity

Verify companies, entities, and agents. GST, MCA21, phone, W3C DIDs.

Enrichment

Transform raw inputs into structured intelligence. IFSC, pincode, company data.

Discovery

Search datasets and capabilities dynamically. Web extract, ground truth.

Intelligence

Subscribe to real-time events. Market data, on-chain activity, competitor changes.

Actions

Execute workflows. Multi-step pipelines, task orchestration, retry logic.

Payments

Native pay-per-use. x402 protocol, USDC on Base, autonomous settlement.

Memory

Persist context across executions. Semantic search and write via Memex.

40 APIS TOTAL
38
production APIs across 14 categories
Browse catalog →
// why yantrix

Agent infrastructure
with governance built in.

AWS runs agents. LangGraph orchestrates them. Yantrix governs them.

Policy enforcement, semantic memory, flow pipelines, orchestration graphs, and autonomous payments — composable layers designed to work together.

Traditional APIs
Yantrix.ai
No spend governance
✓ Policy Engine — every action evaluated
No audit trail
✓ Immutable audit log per agent
Memory lost between sessions
✓ Semantic memory v2 — pgvector
Billing requires humans
✓ x402 autonomous USDC payments
Framework-locked workflows
✓ Flows + Graph — any framework
No M2M contracts
✓ Agent contracts + escrow
// API bundles

Pre-built stacks
for every agent.

Themed collections for common agent use cases. One key works across all of them.

AGENT CORE
Agent Infrastructure
Memory · Relay · Pulse · MCP Registry · Task Orchestration
5 APIs · Agent builders
MOST POPULAR
TRUST
Trust & Verification
GST · MCA21 · Phone · Papertrail · Ground Truth
5 APIs · Verify before acting
INDIA
India Agent Stack
IFSC · Pincode · Indic Engine · Contract Scan
GST · MCA21 — coming soon
4 live · 2 coming soon · India-specific moat
INTELLIGENCE
Intelligence Bundle
Web Extract · Signal · Finance Data · IFSC
4 APIs · Live knowledge
EXECUTION
Execution & Payments
Relay · Settler · Finance Data · Papertrail · Wallet
5 APIs · Autonomous commerce
DEVELOPER
Developer Toolkit
Code Check · Threat Model · Rate Lord · Agent Utility
4 APIs · Safety & infrastructure
// control plane

One control plane
for the entire agent stack.

The architecture that compounds value as workflows and memory accumulate over time.

Agent Action
any framework
Policy Engine
allow / deny
Execution
governed + logged
Audit Trail
immutable record
40
Production APIs
5
Live Services
<50ms
Policy eval
Base
Mainnet
// production ready

Infrastructure built
for production agents.

Yantrix sits between agents and the real world — evaluating actions, tracking spend, and logging decisions so teams can trust agents in production.

Built top-down — payments first, governance second, memory third, orchestration fourth. The only platform where agent autonomy and enterprise accountability coexist.

No credit card required · Cancel anytime after free period

OpenProof signed responses
Every API response cryptographically signed. Verify authenticity on-chain. No competitor has this.
Policy-gated execution + audit trail
Every agent action evaluated by Policy Engine. Every decision stored as an immutable audit record.
Programmable spend policies
Daily limits, per-action caps, and approval workflows. Spend stays within bounds without manual intervention.
5 live governance services
Policy Engine, Memory v2, Flow Pipelines, Orchestration Graph, Gateway. All on Railway. All production-grade.
// governance stack — now live

Four layers of agent control.

Phase 1
Policy Engine
Spend limits, tool permissions, audit trail. Every agent action governed before execution.
policy.yantrix.ai
Phase 2
Semantic Memory v2
pgvector-powered memory. Agents remember across sessions. Data stays in Yantrix.
memex.yantrix.ai
Phase 3
Flow Pipelines
Multi-step pipelines with conditions. Define once. Run autonomously. Context passed between steps.
flows.yantrix.ai
Phase 4
Orchestration Graph
Stateful multi-agent workflows. Parallel execution. Dependency-aware. Developer lock-in layer.
graph.yantrix.ai
Get API key →

No credit card · Wallet required · Pay per call

Production agents.
Built on Yantrix.

Policy-governed execution. Persistent memory. Autonomous payments. Everything production agents need.

Get API key → View documentation

Free for 6 months · No credit card · Wallet required

// get involved

Talk to us.

Found a bug? Want a new API? Building something with Yantrix? We read everything.

Report a bug

Wrong response, broken endpoint, unexpected behaviour.

→ Submit report

Request an API

Need an endpoint or data source we don't have yet?

→ Submit request

Say hello

Partnership, pricing, enterprise, or anything else.

→ Get in touch

Report a bug

Email
Endpoint
What happened?

Request an API

Email
Capability
Use case

Get in touch

Email
Topic
Message