// AIPROCUREMENT.CLUB · v1 · PUBLIC SHOWCASE

Palantir-for-procurement.
Open-source. Audit-verifiable. Multi-agent.

One Postgres-backed kernel for procurement: hash-chained audit log, per-tenant ontology, 10-agent swarm with replayable decisions. Every transition cryptographically verifiable cold. Every spend pattern queryable via 170+ documented APIs. Built to survive a SOC-2 audit on day one.

300
kernel transactions
1 500
audit events on chain
10
specialists in swarm
// SECTION 02

The problem with every procurement stack today

Procurement teams are stuck between two failure modes. The legacy monoliths take 18 months to implement and have no agent layer. The new point tools each cover one slice and shred audit-trail integrity across five vendors. Custom-built always wins for 18 months, then collapses under tech debt.

LEGACY MONOLITHS
Coupa · Ariba
Rigid 12-24 month implementations. €2-6 marginal cost per order. Agent features ship as chat bolt-ons that hallucinate state and cannot be replayed cold.
POINT TOOLS
Tropic · Zip · Pivot
Each one slice of P2P. Audit-trail integrity destroyed across the vendor mosaic. No data-model continuity between intake, sourcing, payment, compliance.
CUSTOM-BUILT
Every CFO ends up here
Wins for 18 months. Then nobody can audit it, the consultants leave, and tech debt accelerates. No primitives — every feature paid for twice.
// SECTION 03

The wedge — three primitives nobody else ships

  • Hash-chained, replayable audit
    Every state transition is hash-chained (RFC 6962 Merkle). Every agent decision is replayable. An auditor verifies a single event against a published root with no platform trust required.
  • Per-tenant ontology that doesn't fork
    Per-tenant ObjectTypes / LinkTypes / Actions extend the global ontology. Workspace branches let you experiment safely. Snapshot-versioned for rollback.
  • 10-agent swarm with persistent memory
    Voyager-style playbook memory (Wang et al. 2023) compounds agent performance over runs. SOC-2 / ESG / OFAC / LkSG / CSRD attestation falls out of the data model.
// SECTION 04

The architecture

One transaction writes domain + audit + outbox atomically. Postgres is the only source of truth. FalkorDB and webhook fan-out are derived — you can drop them and rebuild from the outbox.

┌──────────────────────────────────────────────────────────────────┐
│  Edge / API                                                       │
│  Next.js · /v1/* (170+ routes)                                    │
│  Lucia magic-link + RFC 6238 TOTP · RFC 9457 problem+json         │
└────────────────────────┬─────────────────────────────────────────┘
                         │ tx
                         ▼
┌──────────────────────────────────────────────────────────────────┐
│  SOURCE OF TRUTH                                                  │
│  Postgres 17 · 28 migrations · 60+ tables                         │
│  partitioned audit_events · RFC 6962 Merkle roots                 │
│  outbox table (transactional)                                     │
└──┬───────────────────────────┬─────────────────────────┬─────────┘
   │ outbox                    │ /v1/audit/verify         │ tx
   ▼                           ▼                          ▼
┌──────────┐         ┌──────────────────┐         ┌─────────────────┐
│ BullMQ   │         │ FalkorDB graph    │         │ 10-agent swarm  │
│ workers  │         │ 10 node labels    │         │ Ollama Cloud    │
│ outbox   │         │ 10 edge types     │         │ replayable      │
│ webhooks │         │                   │         │ tool-call audit │
└──────────┘         └──────────────────┘         └─────────────────┘
  • Postgres 17
    ACID source-of-truth. 28 migrations, partitioned audit_events.
  • Outbox-driven projection
    Domain + audit + outbox in one tx. No eventually-consistent gap.
  • FalkorDB graph sidecar
    10 node labels / 10 edge types. Multi-hop supplier-network risk.
  • BullMQ webhooks
    HMAC-SHA256, staircase backoff, 7 attempts → dead-letter.
  • RFC 6962 Merkle audit
    Per-request linear chain, daily Merkle roots, /v1/audit/verify.
  • Lucia + RFC 6238 TOTP
    Magic-link sign-in, hand-rolled TOTP 2FA, marker-based access.
// SECTION 05

What's already shipped — live numbers

Counts pulled live from the seeded prod kernel, refreshed every 60 seconds. Numbers anonymized for public display.

170+
documented /v1/* endpoints
OpenAPI 3.1, Stripe-quality idempotency
28
SQL migrations applied
schema-as-code, custom-migration escape hatch
260
suppliers in graph
6-dimension risk score per row
139
requisitions
catalogue-matched, policy-routed
129
purchase orders
state machine: issued → partial → closed
109
invoices
three-way match against PO + GR
1 500+
audit events on chain
every state transition hash-chained
10
procurement agent specialists
orchestrator + 9, parallel handoff
5
compliance frameworks
OFAC · EU CFSP · UK MSA · LkSG · CSRD
4
fraud detectors
Benford · split-PO · price-drift · consolidation
28
MCP tools exposed
for Claude · Cursor · Cline
10
architecture decision records
every load-bearing choice cited + dated
// SECTION 06

SOTA showcase — anonymized live primitives

Each card below is rendered server-side from real database queries, with supplier names hashed into pseudonyms. Click through to the live (auth-gated) version.

STRATEGIC SOURCING
Kraljic Matrix · auto-segmentation
BOTTLENECKSTRATEGICNON-CRITICALLEVERAGESupplier 1328Supplier 2041Supplier 6769Supplier 5679Supplier 7724Supplier 7134Supplier 9561Supplier 8131Supplier 9614Supplier 8445Supplier 0993Supplier 7666Supplier 7591Supplier 0389Supplier 7352Supplier 5994Supplier 0748Supplier 6132Supplier 3641Supplier 8368Supplier 2024Supplier 7780Supplier 7652Supplier 4225Supplier 9361Supplier 5073Supplier 3961Supplier 0073Supplier 0721Supplier 6639Supplier 5220Supplier 1275Supplier 2695Supplier 1363Supplier 3223Supplier 1918Supplier 8425Supplier 7253Supplier 0385Supplier 2646SUPPLY RISK →SPEND IMPACT →
40 suppliers plotted by spend impact × supply risk. Quadrant assignments drive the sourcing playbook.
VIEW LIVE →
AUDIT PRIMITIVE
PROV-O Hash Chain
EVENT 1813830721:51:36EVENT 233189e721:51:36EVENT 3f9f1c9b21:51:36EVENT 469e819f21:51:36EVENT 5414137321:51:36EVENT 677db2b721:51:36
Last 6 audit events with parent_hash arrows. Every event's hash includes its predecessor — tamper-evident by construction.
VIEW LIVE →
80/20 ANALYSIS
Pareto Spend Curve
Category 267: €44K (cum 13%)Category 415: €35K (cum 24%)Category 831: €32K (cum 33%)Category 001: €32K (cum 43%)Category 481: €28K (cum 51%)Category 209: €27K (cum 59%)Category 323: €27K (cum 67%)Category 597: €24K (cum 74%)Category 848: €24K (cum 82%)Category 033: €23K (cum 88%)Category 988: €20K (cum 95%)Category 338: €18K (cum 100%)80%SUPPLIER RANK →SPEND
Top 12 suppliers by spend with cumulative % overlay. The vital few crossing the 80% line are the negotiation targets.
VIEW LIVE →
AGENT TOPOLOGY
Swarm — orchestrator + specialists
ORCHREQUPOINVOSUPPCOMPLETEDRUNNINGFAILED
Color encodes status (green completed, yellow running, red failed). Edges represent sync handoffs. Live tree shows depth + fan-out per run.
VIEW LIVE →
// SECTION 07

Engineering credibility

10 ADRs documenting every load-bearing architectural choice, with arxiv + RFC + ISO citations. Open-source-ready (repo URL pending public release).

We picked Postgres 17 over MongoDB because financial transactions need ACID. We picked FalkorDB sidecar over pgvector because multi-hop supplier-network queries need a real graph engine. We picked Voyager-style playbook memory because Wang et al. 2023 demonstrated compounding agent performance with persistent lessons.

Read the 10 ADRs →
// SECTION 08

Try it

TRY A 1-CLICK DEMO

Or send your CTO this URL:

https://aiprocurement.club/showcase