frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Open-source KYC plugin for Claude – 95min→27min, £85K→£240/year

https://github.com/vyayasan/kyc-analyst
3•vyayasan•1h ago
Hi HN,

Just launched an open-source compliance plugin for Claude Cowork after seeing fintech teams pay £60K+ for platforms that orchestrate free public data.

UK fintech pilot (30 days, 5 analysts): • 95 minutes → 27 minutes per case • £85K annual platform cost → £240/year (Claude Pro) • Uses only free data: OFAC, UN, EU, Companies House, OpenSanctions

17 mandatory human-in-the-loop checkpoints. No auto-approvals. Deterministic risk scoring (MLR 2017 formulas). MIT licensed.

Launching today because Claude just announced Cowork plugin updates: https://www.linkedin.com/posts/claude-ai_were-introducing-up...

Testing if foundation models can replace compliance middleware for standard workflows (~70% of cases).

Demo slides: https://github.com/vyayasan/kyc-analyst/blob/main/docs/demo-... GitHub: https://github.com/vyayasan/kyc-analyst

Happy to answer questions about LLMs in regulated environments.

Comments

vyayasan•1h ago
Author here.

Quick context on the pilot:

30-day test with 5 compliance analysts at UK fintech doing standard individual onboarding (Know Your Customer checks).

Before: 95 minutes per case (manual searches across OFAC, UN, EU sanctions lists, Companies House, adverse media, PEP databases)

After: 27 minutes per case (Claude orchestrates the searches, analyst reviews at 17 mandatory checkpoints)

Key architectural decision: NO auto-approvals. Every decision requires explicit analyst approval + notes.

Legal team spent 3 weeks reviewing before approving pilot. Main concern was audit trail - solved with immutable markdown logs.

Demo slides show the full workflow: https://github.com/vyayasan/kyc-analyst/blob/main/docs/demo-...

Happy to answer questions about: - The 17 stagegates - How risk scoring works (deterministic, not black box) - Regulatory requirements (FCA/MLR 2017) - What works vs what doesn't

Built this to test if foundation models can commoditize compliance middleware.

felix089•1h ago
This looks interesting, what's your goal with this project?

Show HN: Claud-ometer – See your Claude Code usage, costs, and sessions locally

https://github.com/deshraj/Claud-ometer
1•deshraj•49s ago•0 comments

Show HN: MCP app for viewing/generating 3D Gaussian Splatting in ChatGPT

https://github.com/kstonekuan/splatter-mcp-app
1•kstonekuan•57s ago•0 comments

Sandcastles Made of Bits

https://jsfour.substack.com/p/sandcastles-made-of-bits
2•TheOsiris•1m ago•0 comments

Show HN: VoooAI – natural language to multi-modal AI pipelines

https://voooai.com/
1•hkljjkl•2m ago•0 comments

Every privacy concession in history has been permanent

https://nbtv.substack.com/p/enough-is-enough
1•Cider9986•2m ago•1 comments

Show HN: Ctrl and Click on text to jump to its translation key in VSCode

https://chromewebstore.google.com/detail/i18nkeylocator/nkoandfnjiopdjmhbcnggpeomnmieadi
1•yusufA207•5m ago•0 comments

Xkcd Simulation for Real Packages

https://nesbitt.io/xkcd-2347/
1•michaelw•6m ago•0 comments

New Research Suggests Myopia Could Be Caused by How We Use Our Eyes Indoors

https://www.sunyopt.edu/new-research-suggests-myopia-could-be-caused-by-how-we-use-our-eyes-indoors/
1•gnabgib•6m ago•0 comments

AI and My Crisis of Meaning

https://brids.bearblog.dev/ai-and-my-crisis-of-meaning/
2•mindwok•9m ago•0 comments

Men develop heart disease 7 years before women

https://www.empirical.health/blog/men-vs-women-heart-disease/
1•brandonb•10m ago•0 comments

AI isn't killing SaaS – it's killing single-purpose SaaS

2•JB_5000•11m ago•0 comments

Show HN: WaveGuard – Anomaly detection using wave physics simulation (GPU, MCP)

https://github.com/gpartin/WaveGuardClient
1•waveguard•11m ago•0 comments

Show HN: Jqueue: A job queue that runs on a single JSON file

https://github.com/janbjorge/jqueue
1•jeeybee•17m ago•0 comments

AST-Guided Translation of Natural Language into First-Order Logic with LLMs

https://arxiv.org/abs/2602.13237
1•PaulHoule•17m ago•0 comments

How Long Is the Coast of Britain? (1967)

https://www.jstor.org/stable/1721427
1•Hooke•17m ago•0 comments

Hegseth threatens to blacklist Anthropic over 'woke AI' concerns

https://www.npr.org/2026/02/24/nx-s1-5725327/pentagon-anthropic-hegseth-safety
1•cdrnsf•19m ago•0 comments

Ask HN: Do you use trading AI agents?

1•alphathetaquant•19m ago•1 comments

People-Shaped Problems

https://herbertlui.net/people-shaped-problems/
1•herbertl•19m ago•0 comments

In Pursuit of High-Fidelity GPU Kernel Benchmarking

https://standardkernel.com/blog/in-pursuit-of-high-fidelity-gpu-kernel-benchmarking/
1•matt_d•20m ago•0 comments

Chile's Project Cybersyn (1971-1973)

https://en.wikipedia.org/wiki/Project_Cybersyn
1•alhazrod•20m ago•0 comments

Mac mini plant was planned under the Biden administration

https://appleinsider.com/articles/26/02/24/apples-us-mac-mini-assembly-plant-was-planned-during-t...
3•alwillis•21m ago•0 comments

Cortex 2.0 – The Future of Robotic Intelligence

https://cortex2.sereact.ai
1•pancakenetwork•21m ago•0 comments

GRC Engineering

https://mattgoodrich.com/posts/grc-engineering/
1•mooreds•21m ago•0 comments

Prediction Marketing

https://ayeshaasiddiqi.substack.com/p/prediction-marketing
3•jbegley•23m ago•0 comments

Nerve Compression Syndrome

https://en.wikipedia.org/wiki/Nerve_compression_syndrome
1•Bowes-Lyon•23m ago•0 comments

Show HN: Vigilo – Local audit trail and cost tracker for AI coding agents

https://github.com/Idan3011/vigilo
1•idanr3011•24m ago•0 comments

A TV Transmitter from an STM32

https://hackaday.com/2026/02/24/a-tv-transmitter-from-an-stm32/
2•starkparker•25m ago•0 comments

Citrini Research's '2028 Global Intelligence Crisis': How Worried Should We Be?

https://seekingalpha.com/article/4874066-citrini-researchs-2028-global-intelligence-crisis-how-wo...
2•nis0s•27m ago•1 comments

OpenCode on a Budget – Synthetic.new, Chutes.ai, and Z.ai

https://blog.patshead.com/2026/02/opencode-on-a-budget-synthetic-dot-new-chutes-dot-ai-and-z-dot-...
2•sea-gold•27m ago•0 comments

Ansigpt: C89 Implementation of Microgpt

https://github.com/yobibyte/ansigpt
2•yobibyte•32m ago•0 comments