Agent-Native Data Infrastructure

Every Agent Deserves a Brain.

HatiData is the data layer purpose-built for autonomous AI agents — persistent memory, verifiable reasoning, governed access, through standard SQL.

0sLocal install
0msMemory search (p50)
0MCP tools built-in
0%Cost reduction
In-VPC
CMEK
SOC 2 Ready
PrivateLink
MAS TRM
HIPAA Ready

EXPLAINBUNDLE

One API Call. Full Evidence Chain.

Every artifact carries its complete causal history — which task spawned it, which model decided, what evidence was evaluated, and who approved.

T

mortgage_review_batch_47

kind: compliance_review

A

att_8f2a…c901

agent: compliance-scanner-03

M

claude-sonnet-4-6

routing: cost_optimized

91%

1,847 tokens · $0.008 · 340ms

prompt_hash: sha256:a8f2…

risk_assessment_v2

confidence: 0.94 · validated

94%
3s180ms

Processing latency

420

Custom state tables

NoneFull

Decision explainability

65%22%

False-positive escalation

V2 RUNTIME SUBSTRATE

Three Pillars of Governed Autonomy

Control, transparency, and learning — as platform entities, not application features.

CONTROL

Governed Execution

Branch Contracts enforce visibility modes. Review Gates require evidence bundles before release. Every ReleaseDecision is immutable and auditable.

  • 5 Branch Visibility Modes
  • Human Gate Predicates
  • Immutable ReleaseDecision
  • Lease-Based Ownership
TRANSPARENCY

Causal Explainability

Mandatory lineage records every causal dependency as a DAG edge. ExplainBundle reconstructs the full evidence chain for any artifact in one API call.

  • Mandatory Lineage Edges
  • ExplainBundle API
  • Depth-Limited Graph Traversal
  • Orphan Artifact Detection
LEARNING

Optimization Substrate

Deterministic Replay re-runs past attempts with new models. Shadow Policies evaluate alternative routing without affecting production. Reward Records close the feedback loop.

  • Deterministic Replay
  • Shadow Policy Evaluation
  • Reward Records
  • Confidence Edge Tracking

CUSTOMER STORY

From 42 Tables to Zero Custom Control-Plane

A Series B fintech runs 14 AI agents to process mortgage applications — extracting data, validating compliance, routing to underwriters, and generating decision memos. Before V2, each agent needed its own state table, retry logic, and audit trail. After migrating to HatiData V2, they replaced all of it with platform entities.

3s → 180ms

Document processing latency

42 → 0

Custom state tables eliminated

100%

Decision explainability via ExplainBundle

65% → 22%

False-positive escalation rate

We were spending 40% of engineering time maintaining state machines and retry logic. V2 gave us that time back — and the lineage graph caught two compliance gaps we didn't know existed.

VP Engineering, Series B Fintech

OPERATOR VIEWS

Real-Time Runtime Visibility

SQL-queryable dashboards built on hd_runtime views. No custom monitoring stack required.

Active Leases
compliance-scanner-034:32
doc-extractor-070:48
underwriter-agent-010:12
fraud-detector-023:15
Pending Reviews
income_verification_01compliance@02:14:30
auto_approval_pkgunderwriting@00:45:12
high_value_loan_03risk@00:08:44
Confidence Distribution
≥0.9
847
0.7–0.9
312
0.5–0.7
156
<0.5
52

ZERO TO AGENT MEMORY

From Zero to Agent Memory in 60 Seconds

terminal — hatidata quickstart

# Step 1: Install

$pip install hatidata-agent

# Step 2: Initialize

$hati init

# Step 3: Connect & Query

$from hatidata_agent import HatiDataAgent
$
$agent = HatiDataAgent(host="localhost", port=5439, agent_id="my-agent")
$
$# Store a memory (returns UUID)
$memory_id = agent.store_memory(
$ "User prefers dark mode",
$ memory_type="preference",
$ importance=0.8,
$)
$
$# Search memories by natural language
$results = agent.search_memory("user preferences", top_k=5)
$print(f"Found {len(results)} memories in <5ms")
Memory stored (id: a1b2c3). Found 1 memories in 3.2ms.

No Docker. No signup. No cloud account. Just pip install and go.

APPLICATION PORTFOLIO

What Can You Build?

Six production-ready applications. Each one maps directly to ANDI pillars.

AI Compliance Auditor

CISO / Compliance Officer

Autonomous compliance monitoring with immutable audit trails that regulators trust.

CoT LedgerPer-Agent ABACAgent Memory

Multi-Agent Research Engine

Both (Builder + Buyer)

Multiple AI agents collaborate on research, each with persistent memory and governed data access.

Agent MemoryState BranchingSemantic Triggers

Private Knowledge RAG

AI/ML Engineer

Retrieval-augmented generation that never exposes sensitive data outside your VPC.

Agent MemoryIn-VPC + SQLPer-Agent ABAC

Automated Data Quality Monitor

Data Engineer

AI agents that continuously monitor data quality and explain every anomaly they detect.

Semantic TriggersAgent MemoryCoT Ledger

Real-Time Cost Optimizer

CFO / Head of Data

AI agents that watch your cloud spend and automatically surface optimization opportunities.

Agent MemorySemantic TriggersIn-VPC + SQL

Security Sentinel

CISO / Security Team

AI-powered threat detection with explainable reasoning and zero-trust data access.

Semantic TriggersCoT LedgerPer-Agent ABAC

GET STARTED

Three Paths to ANDI

1

Developer

pip install hatidata → hati init → connect. Local agent memory in 60 seconds. Free forever.

60 seconds

Try the Playground
2

Enterprise

Free audit → In-VPC deployment → 14-day comparison. See what your agents cost before you commit.

14-day pilot

Calculate Savings
3

Strategic

Read the ANDI thesis → Join the Founding 20 → Shape the category with us.

Partnership

Founding 20 Program

COST EFFICIENCY

The Agent Tax Is Real

Legacy warehouses bill 60-second minimums on every micro-query your agents fire.

MetricLegacy WarehousesHatiData
Billing Granularity60-second minimumPer-second
Auto-Suspend60 seconds5 seconds
Compute Markup~300% over cloudYour VPC cost
Agent MemoryExternal vector DBBuilt-in (<5ms)
Audit TrailLog filesHash-chained ledger

70% average cost reduction

Based on median customer workloads. Per-second billing eliminates the 60-second tax on agent micro-queries.

Calculate your savings →

Works with your stack

ClaudeAgent
LangChainFramework
CrewAIFramework
CursorIDE
dbtData
MCPProtocol
OpenAI Agents SDKFramework
AutoGenFramework

Ship Smarter Agents. Start in 60 Seconds.

Persistent memory. Isolated state. Verifiable reasoning. Through standard SQL.