The intelligence layer for open commerce

The Commerce Graph.

India’s open commerce intelligence.

Amazon, Google, Meta and Shopify built intelligence for themselves. Millions of merchants have none. The Commerce Graph connects every commerce signal — and turns it into decisions.

Try it live Take the tour Figures indicative
00 Experience it

Don’t read about the graph. Ask it.

This is the live agent — real Shipsure, serviceability, address and market-trend APIs firing behind every answer. Ask about an order, a pincode, a product.

Open full experience Live · sense.shiprocket.in · rate-limited demo
01 What this is

A shared brain for commerce — not another screen.

Not a dashboard

An intelligence layer

Dashboards show data. This connects context across every system a merchant already uses, and reasons over it.

Not a data lake

A living graph

A lake stores facts. The Commerce Graph understands relationships — buyers, orders, shipments, money and trust.

Not a chatbot

Agents that act

Beyond answers: agents that predict, recommend, execute approved actions, and learn from every outcome.

02 How it works

From signal to memory, in one loop.

01
Signals
Raw events across the lifecycle
02
Schema
A common language for entities
03
Graph
Connected, living relationships
04
Intelligence
Reason · predict · recommend
05
Agents
Execute approved workflows
06
Actions
Change real outcomes
07
Memory
Every outcome makes it smarter
03 The shift

Where commerce is heading.

Today

Fragmented systems

Every platform sees a slice. No system understands the whole business.

The build

Connected intelligence

Signals join into a shared graph that reasons across the full lifecycle.

The horizon

Autonomous commerce

Agents act on trusted context and get better with every decision.

04 Explore

Take the tour.

Why the Commerce Graph

Walled gardens built theirs. Open commerce needs its own.

The signals already exist. What’s missing is the layer that connects them. Open by design. Owned by no walled garden.

Full lifecycle
discovery → checkout → delivery → returns → finance
Open by design
APIs · MCP · CLI · SDKs · webhooks
Network-first
every participant makes it smarter
MOD_01 The Problem

Commerce doesn't have a data problem.
It has a context problem.

Ten systems, ten slices. None understands the business. Hover any surface.

The future of AI in commerce depends on who owns the context layer.

MOD_01.2 Walled gardens vs open commerce

The intelligence gap.

Walled gardens compound intelligence daily — inside their walls. Open commerce needs the whole picture.

Closed intelligence

Built for one ecosystem. Optimizes one platform. Benefits one company.

SearchProduct OrderDelivery RepeatAmazon AI

Amazon learns from Amazon.

Open commerce intelligence

Built across ecosystems. Optimizes the full lifecycle. Benefits every merchant, partner, developer and agent.

MetaGoogleShopifyWhatsApp PaymentsShipmentsSupportFinance Commerce Intelligence

The Commerce Graph learns from all of Indian commerce.

MOD_02 The Graph

From commerce signals to commerce intelligence.

A data lake stores facts. A graph understands relationships. Seven layers, one living model.

1Commerce SignalsRaw events across the lifecycle
2Unified Commerce SchemaA common language for entities
3Commerce OntologyHow commerce actually works
4Commerce GraphConnected, living relationships
5Commerce IntelligenceReason · predict · recommend
6Commerce AgentsExecute approved workflows
7Learning LoopEvery outcome becomes memory

The signal sources

Every source resolves into one shared entity model — some flowing today, some coming soon.

Available todayComing soon

Every source feeds the living graph below — explore how the entities connect, and inspect each one’s exact data points.

MOD_02.2 The living graph

Every entity, connected — like a brain.

The ontology as a living network. Hover to trace. Click any node to inspect its exact data points.

Data available todayComing soon
Hover to trace · click to inspect
MOD_02.3 What the graph knows

Not more data. Better understanding.

A buyer is not a phone number. A shipment is not an AWB. The graph derives intelligence, not records.

Scores simulated · indicative
MOD_03 Intelligence

Imagine if commerce could think.

Each stage generates data, needs a decision, enables an action. Select a stage.

Data available todayComing soon
MOD_03.2 The Engine

It reasons, predicts, benchmarks and acts.

Ask the graph. Every answer carries evidence, actions, impact and confidence.

Ask the Commerce Graph
Suggested prompts
Explain

Understand what happened, and why.

Predict

Forecast what is likely next.

Recommend

Suggest the highest-impact action.

Execute & Learn

Trigger actions, measure outcomes.

MOD_04 Capabilities

AI-powered capabilities with transformative outcomes.

Not one dashboard — a capability layer. Each is grounded in real data, graph entities and executable actions.

Data available todayComing soon
MOD_05 Applications

One intelligence layer. Many applications.

One layer, many users. Pick a persona.

This persona asks
MOD_05.2 Platform

Commerce Intelligence as infrastructure.

APIs, MCP, CLI, SDKs, webhooks. Select an endpoint.

POST
Sample responses · schema and values indicative
MOD_05.3 Developer surfaces

MCP connectors & CLI.

AI agents connect through MCP. Operators and scripts connect through the CLI. Same intelligence, same guardrails.

MCP Server
The Commerce Graph as tools for any AI agent
Coming soon
Connects to
Claude Cursor VS Code Custom agents Enterprise copilots
Tools exposed
get_merchant_360profile, growth score, risk, next best actions
score_order_riskCOD, fraud and address risk before dispatch
recommend_courieroptimal courier with reasons and trade-offs
benchmark_merchantcompare against similar sellers on the network
execute_actionrun an approved workflow and measure outcome
{ "mcpServers": { "commerce-graph": { "url": "https://mcp.commercegraph.example/sse" } } }
CLI
The graph from any terminal or pipeline
Coming soon
$ cgraph merchants top --by growth --limit 3 1. M_urbanveda_001 growth 91 rto HIGH 2. M_saltandpine growth 88 rto LOW 3. M_kettleworks growth 85 rto MED $ cgraph order risk O_44821 cod_risk: low · fraud: 0.08 · address: high → proceed $ cgraph recommend courier O_44821 --explain Courier_B · P(deliver)=0.91 · lane history + address confidence $ cgraph actions run courier-switch --dry-run 512 orders affected · projected RTO 31% → 24% # every command carries evidence and confidence $
SDKs Webhooks Embedded widgets All coming soon
MOD_06 Learning Loop

Every action makes the graph smarter.

Every recommendation becomes an experiment. Every outcome becomes memory.

Worked example

AI recommends a courier switch → merchant accepts → 500 orders tested → RTO falls from 31% to 24% → confidence improves → similar merchants receive the updated recommendation.

Illustrative outcome
MOD_06.2 Network Flywheel

The power of the network.

One merchant is valuable. Thousands compound. Benchmarks sharpen.

More merchantsMore signals Better intelligenceBetter outcomes Stronger graph
BENCHMARK CONFIDENCE
10%
MOD_07 Simulated Future Testimonials

A future we believe is possible.

Illustrative narratives of the outcomes this is designed to create.

Simulated future testimonials · not real customers
Closing Vision

The internet connected information. Commerce platforms connected transactions. The Commerce Graph connects understanding.

From connected signals to a shared intelligence layer to the AI-native operating system for open commerce.

Signals Ontology Graph Intelligence Agents Actions Memory Network Flywheel

Living Graph — Admin

local edits
Entity
Data points — click a status to toggle live ↔ coming soon
Saved in this browser · Export to make it permanent in the HTML