All 8 systems.

FiskalLink AI

shipped

Ministry of Finance

Tagging a region's budget to a policy theme, line by line.

It turns a plain-language policy theme into a reviewed, traceable analysis of a large budget.

1.83M

budget records reconciled

USD 224.4B

public expenditure covered

667 : 1

ledger rows per AI judgment, across 2,741 judgments

• Reconciles 1.83 million budget records covering USD 224.4 billion of public expenditure
• Turns a five-expert-day analysis into a 20-minute workflow while preserving human review
• Deployed on AWS Lightsail Jakarta with Cloudflare edge to satisfy data residency

Open the record

Klaussa

in production

Seeded legal-tech company

Indonesian law is scattered across 37 government sites, much of it only as scans.

It answers legal research questions from Indonesian regulations, with the source and citation trail attached.

297K

regulations acquired from 37 sources

3.48M

vectors indexed

2.57M

citation edges at ~99% precision

• Indexes 300,000+ Indonesian legal documents on self-hosted Qdrant
• Serves roughly 100 daily users
• Custom OCR, a clause-aware chunker and an in-text citation-graph algorithm

Open the record

Paperbase

built, no external users

Own research infrastructure

An agent cannot tell you whether your idea is already published.

An index of AI research that keeps itself current, which an agent queries through nine tools over MCP. The pipeline works down a venue roster, pulls each paper's metadata, fetches the arXiv PDF and extracts the full text, so a phrase buried in a body paragraph is findable. A paper outside the index comes back as a visible stub that one call will ingest.

159,815

AI papers indexed

106,263

of them with full text extracted

9

query tools exposed over MCP

Own research infrastructure. It answers this author's agents over MCP and is not offered to anyone else.

Open the record

Klaussa Retriever

built, no users recorded

Klaussa · retrieval research

The law that answered the question sat at rank 14.

Indonesian law has whole families of near-identical titles, and a general encoder ranks them on how similar the wording looks, so the one that actually governs can land at rank 14. A 568M-parameter retriever fine-tuned on this corpus puts it first, and answers in about 40 ms on an ordinary CPU against about 500 ms for the hosted API.

14 → 1

rank of the governing document, this query

12.5×

faster per query, ~40 ms local against ~500 ms API

0.496

nDCG@10 on the agent register against 0.413, n = 803

Measured rather than adopted: scored on two registers of the Klaussa pool, n = 803 and n = 545, against the hosted embedding API that serves production. It is not one of the systems listed as shipped or in production, and no user figure is recorded for it.

Open the record

ResepCerdas

deployed, no users recorded

Faculty of Pharmacy, Universitas Islam Sultan Agung

A prescription is written in Latin shorthand, and the patient has to take it correctly.

It turns a photographed prescription into corrected drug lines, then checks each line with 34 deterministic administrative, pharmaceutical, and clinical rules. No screening verdict comes from a language model.

34

deterministic checks: 10 administrative, 10 pharmaceutical, 14 clinical

19 / 19

fields read on both benchmark fixtures, clean and degraded — synthetic handwriting, not a doctor's own

2.2 s

extraction latency on the clean fixture, 2.3 s on the degraded one

Evaluated on synthetic prescriptions; not yet cleared for real patient data.

Open the record