AI-native SaaS for regulated industries · Globally

Software that
delivers
outcomes.

AI-native SaaS for regulated industries — Integrator (ETL and system integration), Optimizer (code and process performance), DocuGen (documentation generator from code repositories), SecureAudit (security review generator from code repositories) — built on the Anthropic Claude API. CCA-F certified (Claude Partner Network) in Claude solution architecture.

Our services Book a pilot
4
AI-native services
CCA-F
Claude certified
Global
Service coverage
OSS
pyvar.com
CCA-F · Claude Partner Network
Basel III/IV
FRTB IMA/SA
MiFID II
EMIR
GDPR compliant
Services

Four AI-native services,
one platform.

AI-native SaaS for regulated industries, all built on the Anthropic Claude API: ETL and system integration generated from your schemas and API descriptions, code and process performance, and documentation and security reviews generated from your code repositories.

01
Integrator
Software Integration as a Service (SIaaS)
ETL and system integration, generated by Claude from your schemas, API descriptions and sample data: Python ETL scripts, API connectors and webhook handlers in seconds — and nothing you upload is stored. Live now in alpha.
ETL scriptsAPI connectorsWebhooks
Explore Integrator →
02
Optimizer
Software Optimization as a Service (SOaaS)
Code and process performance. Claude profiles your codebase and workflows to improve performance, efficiency and code quality — from algorithmic complexity down to database queries.
PerformanceEfficiencyCode quality
Explore Optimizer →
03
DocuGen
Software Documentation as a Service (SDaaS)
A documentation generator built from your code repositories — automated, accurate and always current, kept in sync with every commit, including compliance-ready doc sets for regulated environments.
AutomatedAccurateAlways current
Explore DocuGen →
04
SecureAudit
Software Security Review as a Service (SSRaaS)
A security review generator built from your code repositories, covering vulnerability scanning, credential detection and remediation — delivered as a pull request, not a PDF.
Vulnerability scanningCredential detectionRemediation
Explore SecureAudit →
Consultancy · By arrangement
Consultancy for AI-native software in regulated industries
A tailored engagement, not a product: scoped with you, one to one. Tell us what you are building and the rules you work under, and we will reply by email to agree whether and how we can help.
Discuss your project What it covers → consultancy@fibtec.co.uk
Tailored
Every engagement
1:1
Scoped with you
Email
How to start
Also by Fibtec · Open source
pyvar.com — Financial Risk Computation Platform
385 regulatory-grade financial and risk functions across 8 domains, exposed as a REST API. Built on NumPy, Numba and the Anthropic Claude API. Free, open source (Apache-2.0 licensed), and separately hosted at pyvar.com.
Visit pyvar.com ↗ View on GitHub ↗
385
Functions
8
Domains
OSS
Open source
Also by Fibtec · Beta release
tengrade.co.uk — Anonymous, Auditable Exam Grading
TenGrade grades open-ended exam answers anonymously and auditably. It does not hire, rank, or decide — that responsibility stays with you. Currently in beta release, separately hosted at tengrade.co.uk.
Visit tengrade.co.uk ↗ Claude Code plugin on GitHub ↗
Anonymous
Candidates
Auditable
Every result
Beta
Status
Also by Fibtec · Open source
claude-docker-public — Containerized Sandbox for Claude Code
A Docker setup for running Claude Code with the container itself as the security boundary — plugins baked into the image at build time, zero network calls at container start, no reliance on a permission prompt. Apache-2.0 licensed, and the same setup we use to build pyvar.com.
View on GitHub ↗
41
Plugins baked in
0
Live network calls
OSS
Open source
How we work

Every engagement is a triangle.
Savings, speed, accuracy.

Across all four services we track the same three axes — savings, speed and accuracy — each normalised so 0 is no improvement and 1 is the theoretical best. Committing to a target triangle up front and overlaying what's actually delivered gives an honest, comparable read on every engagement: the bigger the triangle, the more efficient the outcome.

Iron Triangle: savings, speed and accuracy Illustrative expected-vs-actual comparison, inverted so bigger reads as better. Expected: savings 0.65, speed 0.70, accuracy 0.75. Actual: savings 0.85, speed 0.90, accuracy 0.90. Values are normalised 0 (no improvement) to 1 (theoretical best), plotted on three axes 120° apart. Savings Speed Accuracy
Expected — committed at kickoff
Actual — delivered

This is an illustrative worked example, not a specific client's measured result — it shows how the model reads, not a guaranteed outcome. Any single axis at 0.2 or below is flagged regardless of the combined score: the triangle is the headline, the three axes are the audit.

Expected
Savings0.65
Speed0.70
Accuracy0.75
TES 9.1/10
Actual
Savings0.85
Speed0.90
Accuracy0.90
TES 9.9/10
About Fibtec

Financial engineers,
now AI-native.

Fibtec is an AI-native software company building outcome-focused Claude applications for regulated industries — financial services and beyond.

Every service in our suite is built on the Anthropic Claude API, and our team holds CCA-F certification (Claude Partner Network) in Claude solution architecture. We also built pyvar.com, our open-source risk computation platform, to give the financial engineering community free access to institutional-grade risk functions, and TenGrade, our anonymous exam-grading platform, currently in beta release.

Start a pilot pyvar.com ↗ tengrade.co.uk ↗
Built on
The Anthropic Claude API
Certified in
CCA-F · Claude solution architecture
Focused on
Regulated industries
Also built
pyvar.com — open source risk platform
Also built
tengrade.co.uk — anonymous exam grading (beta)
Get in touch

Ready to see outcomes?

Tell us what you're building and we'll show you how Claude can ship it — integration, performance, documentation or security.

info@fibtec.co.uk

CCA-F certified (Claude Partner Network) · built on the Anthropic Claude API