SHANGO TRUSTGRAPH — Now in Early Access

Trust infrastructure for autonomous agents.

Shango TrustGraph gives AI agents a simple way to verify claims, check permissions, approve actions, score risk, verify deliverables, and create audit receipts before they act.

curl -X POST https://api.odemicroservices.llewellynsystems.com/v1/trust/score

Live Response Preview

POST /v1/trust/score — 200 OK — 38ms
87/ 1000100
confidence: 94%
TRUST RECEIPT
receipt_idtreceipt_01J4X8KZ3WNQ5BV0GFPRTMYD7C
entity_idusr_7f3a9b2e
entity_typeuser
score87
confidence0.94
scored_at2026-05-14T09:22:11.432Z
flags[]
HMAC-SHA256 signed

Dimension Breakdown

Identity Trust
verified91
Behavior Trust
normal83
Permission Trust
authorized88

Evidence Graph

43Nodes
61Edges
egraph_9fk2Graph ID

Context

platform:"marketplace"
transaction_value:"4500"
region:"us-east-1"

Eight Dimensions

Every angle of trust. One API.

TrustGraph scores across eight independently weighted dimensions. Use all eight or cherry-pick the ones your workflow demands.

Identity Trust

Verify who they are. Cross-reference signals across identity graphs.

Source Trust

Where did this come from? Trace provenance and chain of custody.

Claim Trust

Is this true? Score factual accuracy against verified knowledge graphs.

Content Trust

Is this authentic? Detect synthetic media, AI-generated text, and manipulations.

Transaction Trust

Is this deal legit? Assess risk, counterparty history, and settlement confidence.

Behavior Trust

Does this pattern hold? Identify anomalies across historical behavioral sequences.

Permission Trust

Are they authorized? Verify delegation chains, scopes, and access rights.

Deliverable Trust

Did they deliver? Close the loop with evidence-backed completion verification.

AI agents can write, browse, summarize, and execute.

But they still struggle with the questions that matter before action:

  • Is this actually true — or just plausible?
  • Is this action safe and authorized?
  • Is this entity who they claim to be?
  • Will this content get you banned or sued?
  • Has the deliverable actually been completed?
  • Can you prove this decision was sound?
  • Who takes the blame when the agent is wrong?

Shango answers those questions with structured trust decisions and auditable receipts.

Product

One API. Six trust decisions.

01

Claim verification

Stop your agents from publishing unverifiable facts.

02

Publish-risk checks

Know before you post whether it will get flagged, rejected, or banned.

03

Contactability checks

Run outreach at scale without getting marked as spam or worse.

04

Entity resolution

Verify creators, athletes, and brands are real before you pay them.

05

Action approval

Give your autonomous agents a brake pedal, not just an accelerator.

06

Deliverable verification

Release payment only after the work actually meets requirements.

How It Works

Three steps to verified trust.

01

Submit

POST an entity, claim, or transaction to the TrustGraph API with the dimensions you want scored.

02

Score

Our graph engine traverses evidence nodes, weighs signals, and computes a 0-100 composite trust score with per-dimension breakdowns.

03

Receipt

Receive a signed, immutable trust receipt with score, evidence graph, confidence interval, and a unique receipt ID for your audit trail.

Partner Program

Route trust checks to Shango. Earn 40%.

Your agent brings the workflow.

We handle the trust layer.

You share the revenue.

Join partner program

Trusted Infrastructure

Used across the Llewellyn Systems AI agent network.

SHANGO
ETHER
APEIRON

Ready to score trust at scale?

Start with 100 free checks. No credit card required.

Start Free — 100 trust checks