OPERATING SYSTEM FOR AUTONOMOUS AGENTS
OPEN SOURCE AGENT CLEARINGHOUSE

Govern AI Agents.
Prevent Runaway Spend.

/ The open-source control plane for autonomous agent fleets. Enforce PostgreSQL row-locked spending limits, sandbox MCP tools, and verify every step in an immutable SHA-256 audit ledger.

VIEW ON GITHUB
■■■ROW-LOCK BUDGETS
■■■MCP SANDBOXING
■■■MIT LICENSED
TOPOLOGY: DISTRIBUTED RUNTIME6/6 SUBSYSTEMS ARMED
CLEARINGHOUSEAGENTOS KERNEL
ACTIVE SUBSYSTEM: BUDGET WALLET(Row-Lock Clearinghouse)
<1.1ms lock latencyACTIVE
SYSTEM ARCHITECTURE DIAGRAMATOMIC VERIFICATION
01 INPUT
API PROMPT
WEBHOOK TRIGGER
TASK PAYLOAD
02 KERNEL
ROW-LOCK CHECK
SSRF PROXY
ED25519 SIGN
03 RECEIPT
PASSED [200 OK]
SHA-256 HASH
LEDGER SYNC
TRANSACTION GUARANTEE:POSTGRESQL WAL ATOMIC COMMIT

Deterministic Control.
Zero Token Leakage.

Autonomous agents execute without human supervision only when guarded by deterministic rails. AgentOS intercepts every tool invocation, enforces PostgreSQL row-level locks on your balance, and signs every step with Ed25519 cryptography.

<1.1ms
Lock Latency
0.2ms
Injection Filter
100%
Verifiable Trail
ARCHITECTURAL SPECIFICATION

Engineered For Reliability.
Built For Autonomous Scale.

Six architectural pillars guarantee deterministic execution across your entire agent fleet.

01 / CLEARINGHOUSE

Row-Lock Spend Limits

Prevents infinite recursive agent loops from draining API balances. Every token expenditure executes within an atomic PostgreSQL transaction lock.

DAILY LIMIT: $50.00$14.20 CONSUMED
STATUS: NORMALCIRCUIT BREAKER: ARMED
02 / PERMISSIONS

MCP Tool Sandboxing

Enforce Model Context Protocol compliance. Every tool payload is validated against strict JSON schemas with built-in SSRF protection against private cloud metadata.

REQ: GET /metadata/169.254[BLOCKED]
REASON: RFC 1918 PRIVATE EGRESS FORBIDDEN
POLICY: RING-0 RESTRICTED
03 / ORCHESTRATION

DAG Workflow Engine

Chain tasks across specialized agents in directed acyclic graphs. Supports checkpointing, state recovery, and human approval gates for critical decisions.

TRIGGERPROCESSAPPROVE
OUTBOX: 0 PENDINGSTATE: CONSISTENT
04 / SECURITY

Ed25519 Machine Keys

Every autonomous agent is provisioned with an asymmetric keypair. Requests and tool executions are signed at source, preventing agent impersonation.

PUBLIC KEY:
ed25519:pk:9f28e01ca7b82c3...
SIG: RFC 8032 VERIFIEDPASS
05 / PERSISTENCE

pgvector Semantic Memory

Episodic and working memory backed by native PostgreSQL vector embeddings. Query historical conversation context with low-latency cosine distance searches.

QUERY: "previous stripe order"0.912 SIM
INDEX: HNSW (m=16, ef_construction=64)
RECALL LATENCY: 2.3ms
06 / VERIFICATION

SHA-256 Audit Chaining

Immutable, append-only ledger of every decision, prompt, and tool call. Consecutive entries cryptographically bind to prior hashes, eliminating tampering.

HEAD: #42hash: 8f40a...c219VALID
PREV: #41hash: a4c8e...7f21LINKED
DEPLOYMENT PIPELINE

Launch In Minutes.
Operate With Certainty.

No complex distributed orchestration boilerplate. Provision your first guarded agent in under four commands.

START FREE WORKSPACE
01STEP ONE

Connect Providers

LINK ANY OPENAI-COMPATIBLE MODEL PROVIDER, ANTHROPIC, OR LOCAL LLM INSTANCES IN SECONDS.

02STEP TWO

Configure Rails

DEFINE HARD SPEND CAPS, HOURLY TOKEN LIMITS, AND SSRF-PROTECTED TOOL WHITELISTS.

03STEP THREE

Deploy Agents

PROVISION AUTONOMOUS WORKFLOWS WITH ASYMMETRIC ED25519 MACHINE IDENTITY PAIRS.

04STEP FOUR

Audit Real-Time

STREAM REAL-TIME EXECUTION TELEMETRY AND VERIFY IMMUTABLE SHA-256 EVENT CHAINS.

REAL-WORLD USE CASES

Production Showcases.
Deterministic In The Real World.

See how engineering and operations teams deploy autonomous agent fleets without compromising safety or budget ceilings.

ENGINEERING INFRASTRUCTURE

Automated PR Triage & Security Review

Autonomous code review agents inspect pull requests, run linting suites, test for common vulnerability patterns, and post inline comments within budget limits.

GITHUB MCPCODE REVIEW AGENTSLACK DISPATCH
ATOMIC ROW-LOCKS
100% BUDGET ENFORCED
MCP SANDBOXING
ISOLATED EGRESS
FINANCIAL OPERATIONS

Ledger Reconciliation & Invoicing

Agents verify Stripe billing webhooks against PostgreSQL database orders, generate discrepancy reports, and trigger settlement actions with zero human drift.

STRIPE WEBHOOKRECONCILIATION AGENTACCOUNTING LEDGER
CIRCUIT BREAKER
ZERO OVER-BUDGET RUNS
POSTGRES WAL
TRANSACTIONAL OUTBOX
SECURITY OPERATIONS

Continuous Threat Sentinel & Prompt Guard

Payload inspection models scan incoming agent instructions for prompt injections, jailbreak vectors, and credential leakage before execution reaches tools.

INCOMING PAYLOADLAYER 8 GUARDEXECUTION ROUTE
REAL-TIME DLP
0.2ms FILTER LATENCY
AUDIT LEDGER
SHA-256 VERIFIED
DEVELOPER TOOLCHAIN

Built For Engineers.
Idiomatic SDKs & CLI.

Integrate hard spend limits and cryptographic audit trails with a single library call.

import { AgentOS } from "@agentos/sdk";

const agent = await AgentOS.agents.create({
  name: "sentinel-pr-reviewer",
  model: "gpt-4o",
  spendLimit: { daily: 25.00, perRun: 2.50 },
  tools: ["github-mcp", "slack-webhook"],
  capabilities: ["read_memory", "write_memory"],
});

const execution = await agent.execute(
  "Review open PR #142 for potential secret leakage"
);

console.log("Cost:", execution.cost); // 0.08 USD
console.log("Audit Hash:", execution.auditHash); // "a4c8e...7f21"
AgentOS Interactive CLI Simulator
$agentos verify-audit
╔══════════════════════════════════════════════════════════════════════╗
║                    Layer 6 Ledger Verification                       ║
╚══════════════════════════════════════════════════════════════════════╝
✓ Ledger Cryptographically Verified (SHA-256 Chaining Unbroken)

Total Events Chained: 42
Genesis Hash: 0000000000000000000000000000000000000000000000000000000000000000
Head Digest: 980c9339762ed5bd6ef2596ab847a9ecfa03c1ef7e69a03975a1d95392e21b76
Tamper Status: 0 Alterations Detected (Ledger Verified)
RUNTIME SPECIFICATION

Complete 8-Layer Architecture.
Inspect Every Protocol.

Explore the deterministic execution layers powering AgentOS from hardware keypairs to clearinghouse ledgers.

LAYER 1: ASYMMETRIC IDENTITY

Hardware machine keys with Ed25519 signing

Every autonomous agent is provisioned with a cryptographic Ed25519 keypair upon creation. Actions and transactions are signed client-side by the agent's private key, enabling tamper-evident machine identity without transmitting long-lived credentials over the network.

SIGNATURE ALGORITHM

Ed25519 (RFC 8032)

TOKEN LIFESPAN

15-Minute Revocable JWTs

ZOMBIE DEFENSE

Redis-backed JTI blocklist

ATTESTATION LATENCY

< 2ms

machine_attestation.ts
import { AgentOSClient } from "@agentos/sdk";

const client = new AgentOSClient({ apiKey: process.env.AGENTOS_KEY });

// 1. Register agent with asymmetric identity
const agent = await client.agents.create({
  name: "Sentinel-Financial-01",
  vertical: "finance",
  generate_keypair: true,
  hard_budget_limit: 500.0,
});

// 2. Sign autonomous intent with agent private key
const signature = await signIntent(agent.private_key_pem, {
  action: "execute_rebalance",
  timestamp: Date.now(),
});

// 3. Cryptographic attestation verification
const verification = await client.agents.verifySignature(
  agent.id,
  actionPayload,
  signature
);
console.log("Attestation Verified:", verification.verified);
TRANSPARENT VALUE

Predictable Rails.
Scale At Your Pace.

Start with MIT open-source self-hosting. Upgrade to cloud runtime when your team needs hosted clearinghouse infrastructure.

COMMUNITY & OSS
$0
FOREVER FREE · MIT LICENSE

Ideal for individual developers building local prototypes and testing MCP tool integrations.

START FOR FREE
Up to 3 active agent runtimes
Local MCP tool registry
In-memory event audit log
Docker Compose self-hosting
Community Discord support
TEAM & GROWTHPOPULAR
$49 /month
BILLED MONTHLY

For engineering teams deploying autonomous production agents with hard financial rails.

START 14-DAY FREE TRIAL
Up to 25 active agent runtimes
Row-lock budget clearinghouse
Model Context Protocol sandboxing
Cryptographic SHA-256 audit ledger
Real-time SSE event telemetry
Priority email & developer support
ENTERPRISE
Custom
CUSTOM ANNUAL CONTRACT

For organizations requiring dedicated cloud VPC clusters, on-prem runtimes, and custom SLAs.

CONTACT SALES
Unlimited agent runtimes
Dedicated VPC or on-prem deployment
Hardware Security Module (HSM) keys
Custom security ring policies
99.95% availability SLA guarantee
Dedicated solutions architect
ECOSYSTEM INTEGRATIONS

Connect Everything
You Already Use.

Compatible with any OpenAI-format API endpoint, local Ollama/vLLM server, and Model Context Protocol (MCP) tool server.

LLM PROVIDER
OpenAI
STATUS:NATIVE
LLM PROVIDER
Anthropic Claude
STATUS:NATIVE
LLM PROVIDER
Google Gemini
STATUS:NATIVE
LLM PROVIDER
DeepSeek
STATUS:NATIVE
OPEN MODEL
Meta Llama
STATUS:SUPPORTED
INFERENCE
Groq LPU
STATUS:SUPPORTED
DATABASE
PostgreSQL
STATUS:ROW-LOCK KERNEL
CACHE & QUEUE
Redis
STATUS:SUPPORTED
TOOL PROTOCOL
GitHub MCP
STATUS:VERIFIED
TOOL PROTOCOL
Slack Webhook
STATUS:VERIFIED
PAYMENTS
Stripe API
STATUS:VERIFIED
BROWSER AGENT
BrowserOS
STATUS:VERIFIED

Deploy Your First Agent
With Verifiable Rails Today.

No credit card required. MIT licensed open-source runtime. Full feature access. Live in under two minutes.

■■POSTGRES ROW-LOCKS
■■ED25519 MACHINE KEYS
■■MIT OPEN SOURCE