Products we build,
run, and own.
The lab is where Skyletica turns research into software people use every day. Four products are already in the hands of customers, covering AI agents, financial document forensics, AI accounting, and personal data protection.
How the Lab Thinks
From software house
to product lab.
Skyletica began as a software house. The lab is what we became once we started building, operating, and answering for our own products.
Future-first research & anticipation
We pick problems by where regulation, AI capability, and operating cost are heading, not only where they are today. That is how a privacy vault and a tax-aware ledger ended up on the roadmap before customers asked.
Lab-grade execution & iteration
Every idea starts as a working prototype against real data, then earns its way to production through evaluation, hardening, and feedback from real users. Ideas that do not survive that process are retired early.
Scalable impact & product ownership
We build, run, and support our own products. Owning the pager is what keeps the engineering honest, and it is the same discipline we bring to the systems we build for clients.
Portfolio
What the lab
has shipped.
Each product started as a problem we saw repeatedly in client work. Here is the problem, what we built, and where it stands today.
Enableo
LiveNo-code AI agent platform
A no-code platform for building, deploying, and operating AI agents that hold real conversations and complete real workflows, grounded in your own documents and connected to your systems.
The problem
Teams wanted AI agents that do more than answer FAQs, but every serious build needed engineers for retrieval, memory, routing, and integrations.
What we built
- Visual workflow builder with trigger, agent, HTTP, memory, and data transform nodes
- Document knowledge base with locally computed multilingual embeddings
- Multi-session memory, embeddable web widget, and developer API keys
- Agent pools with round-robin and A/B routing, plus ticketing and human escalation
- Bring-your-own LLM keys across OpenAI, Anthropic, and Gemini, encrypted with AES-256
- Usage, cost, and conversation analytics with white-label deployment
FinGuard
LiveFinancial document forensics
AI-powered screening, reconciliation, and forensic analysis of identity and financial documents that detects tampering, inconsistencies, and fraud signals before a credit or compliance decision is made.
The problem
Lenders and auditors were reviewing KTP, NPWP, payslips, and bank statements by eye, and edited PDFs were getting through.
What we built
- Staged worker pipeline: ingest, classify, extract, evaluate, and profile
- Calibrated OCR extraction for Indonesian identity and financial documents
- Rule evaluation across five detection layers, from metadata to arithmetic
- Detection of metadata edits, template doctoring, and figures that do not reconcile
- Per-applicant risk profile with a complete, reviewable audit trail
Nerati
BetaAI accounting for Indonesian businesses
Double-entry bookkeeping that balances itself, with full Indonesian tax compliance, ten financial statements, and Nera, an AI accountant that answers from your posted journals.
The problem
Indonesian SMEs juggle spreadsheets, a separate tax tool, and an accountant they see once a quarter. None of them talk to each other.
What we built
- Automatic double-entry journals from invoices and bills, on accrual or cash basis
- Ten financial statements and industry chart-of-accounts templates
- PPN and e-Faktur, SPT Masa, and PPh 23, 22, 26, and 4(2) withholding with Coretax export
- Nera, an AI accountant grounded in your ledger that answers in Bahasa Indonesia
- Anomaly insights at four severity levels and a 13-week cash forecast
SkyHold
LiveData privacy vault
An HSM-backed vault that encrypts and tokenises personal data, enforces retention, and keeps data in the region the law requires, so applications never hold raw identifiers.
The problem
Personal data protection laws now carry real penalties, yet NIK numbers, emails, and IDs were sitting in plain text across application databases.
What we built
- Envelope encryption with a per-record data key and an HSM-held master key
- Tokenisation of NIK, email, phone, and ID numbers
- Retention policies with ISO 8601 durations and delete or archive actions
- Automated key rotation and a tamper-evident audit trail
- Region-pinned vaults and scoped API keys, as cloud SaaS or on-premises
Inside the Products
Real screens,
not mock-ups.
A look at the interfaces customers use today.
Enableo
Visual agent workflow builder
Nerati
Nera answering from posted journals
Enableo
Document knowledge base
Nerati
13-week cash forecast
Enableo
Ticketing and human escalation
Nerati
Indonesian tax compliance
Research Tracks
Where the next
products come from.
Six areas of ongoing research. Most have already shipped into a product; the rest are being prototyped with partners.
Multi-agent orchestration
Our proprietary orchestration engine coordinates multiple AI agents, tools, and memory across a workflow graph, with routing, retries, and human hand-off built in.
Shipped in EnableoGrounded AI for regulated data
Assistants that answer only from systems of record, cite where an answer came from, and decline when the data is not there. Built for finance, tax, and compliance work.
Shipped in NeratiDocument intelligence & forensics
Classification, calibrated extraction, and layered rule evaluation that turn scanned identity and financial documents into structured, verifiable evidence.
Shipped in FinGuardPrivacy engineering & data sovereignty
Envelope encryption, tokenisation, retention enforcement, and region pinning that let applications use personal data without ever holding it in the clear.
Shipped in SkyHoldDistributed intelligence systems
Architectures for transparent, secure, and scalable AI that runs across distributed environments, from cloud regions to on-premises deployments.
Active researchAI evaluation & assurance
Bespoke evaluation sets, adversarial testing, and production monitoring that tell you whether an AI system behaves before and after it goes live.
Used across every productFrom Idea to Product
Four stages,
no shortcuts.
Anticipate
Spot a problem that regulation, cost, or new AI capability is about to make urgent, and validate it with the people who feel it.
Prototype
Build a working slice against real data within weeks. Measure it, show it to users, and decide whether it deserves to continue.
Harden
Add multi-tenancy, security, audit trails, evaluation harnesses, and the operational tooling a production system needs.
Ship & evolve
Release to real customers, run it ourselves, and keep improving from usage data and support conversations.
Technology
The stack behind
the products.
Proven, well-supported technology chosen for the long run, the same stack we use in client engagements.
Frontend & mobile
- React
- Next.js
- TypeScript
- Tailwind CSS
- React Native
Backend
- Node.js
- NestJS
- Express
- Python & FastAPI
- Go
Data
- PostgreSQL
- MongoDB
- ClickHouse
- TimescaleDB
- Redis
AI & ML
- OpenAI
- Anthropic Claude
- Google Gemini
- Local embeddings (bge-m3)
- scikit-learn & Prophet
Cloud & operations
- Kubernetes
- Terraform
- AWS & GCP
- S3-compatible storage
- Prometheus, Grafana & ELK
Security
- HSM key management
- AES-256 encryption
- Role-based access control
- Scoped API keys
- Audit trails
Integrations
- Coretax DJP & e-Faktur
- Payment gateways
- WhatsApp Business API
- REST, webhooks & SSE
- ERP, CRM & HRIS
Work With the Lab
Four ways
to partner.
Whether you want to adopt a lab product or build the next one with us.
Co-development
Your domain expertise, our product engineering. We build a new product or module together, with clear ownership terms from day one.
Pilot programmes
A time-boxed pilot of a lab product on your real data, with agreed success measures and a straight answer at the end.
White-label & licensing
Run a lab product under your brand or inside your own infrastructure, with on-premises and region-pinned options where available.
Research partnerships
Joint work on problems without an off-the-shelf answer yet, from agent orchestration to privacy engineering.
Looking for a custom build rather than a lab product?
View our solutionsTrusted By
Organisations that
work with Skyletica.
Industries we have delivered for
FAQ
Questions about the lab.
What is the Skyletica Labs Innovation Lab?
It is the part of Skyletica Labs that designs, ships, and operates our own products. Everything in the portfolio on this page was conceived, engineered, and is run in production by the lab, which is why we can speak plainly about what each product does and does not do.
Which products has the lab built?
Four: Enableo, a no-code AI agent platform; FinGuard, financial document screening and forensics; Nerati, AI accounting and tax compliance for Indonesian businesses (currently in beta); and SkyHold, an HSM-backed data privacy vault.
Can we co-develop a product or run a pilot with the lab?
Yes. We run co-development programmes where your domain expertise meets our engineering, time-boxed pilots on a lab product with your real data, and research partnerships on problems that do not yet have an off-the-shelf answer.
Can lab products be white-labelled or deployed on-premises?
Enableo supports white-label deployment, and SkyHold is available as cloud SaaS or on-premises with region-pinned vaults. For other products we discuss licensing and deployment options case by case.
How does the lab keep AI output trustworthy?
Our products ground AI in systems of record rather than the open internet: Nera answers from posted journals, Enableo agents answer from your documents, and FinGuard explains every flag with the rule that raised it. We pair that with evaluation sets, guardrails, and audit trails.
Let's Build the Future Together
Ready to transform your organization with AI? Get in touch.