Five Ways to Access European Bank Data in 2026 and What Each Costs

·1 views

Fintech developers building apps in Europe can access bank transaction data through five main routes under the PSD2 directive, each carrying different costs and complexity. Direct AISP registration offers the most control but requires €4,000–€16,000 or more annually in certificates and legal fees, plus significant engineering time per bank. Aggregator platforms like Plaid, TrueLayer, and Yapily simplify integration to days but charge $0.50–$2.00 per connected account monthly, which scales steeply with user growth. Nordigen, once known for a generous free tier, has shifted pricing focus toward enterprise clients following its acquisition by GoCardless. Middleware API providers represent a fifth option, abstracting away certificate requirements entirely and offering simple API-key access for developers who want to avoid regulatory overhead.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

How to Build a Production-Ready Feedback System With TypeScript and PostgreSQL

A developer tutorial published on DEV Community outlines how to build a structured feedback workflow using TypeScript, Express, and PostgreSQL, going beyond a simple feedback button to create a fully operational system. The guide models a feedback lifecycle with defined statuses such as new, acknowledged, planned, and resolved, ensuring every item has an owner, a next-action timestamp, and a reply path. It proposes a four-part architecture where PostgreSQL serves as the source of truth and a transactional outbox pattern reliably delivers notifications to chat tools, email, or issue trackers. The tutorial provides SQL schemas for three tables — feedback, feedback_event, and feedback_outbox — along with guidance on capturing diagnostic context without collecting sensitive user data. The design is intended to prevent feedback from going unactioned by making inactivity queryable and routing replies back through the original channel.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Saga Pattern, Outbox and Idempotency: Keeping Consistency in Distributed Systems

As software architectures shift from monolithic relational databases to microservices, NoSQL stores, and event queues, traditional ACID transactions are no longer sufficient for ensuring data consistency. The Saga Pattern addresses this by breaking large operations into smaller local transactions, each paired with a compensating action to reverse changes if a later step fails. Two orchestration approaches exist: choreography, where services react independently to events with low coupling but implicit flow, and orchestration, where a central coordinator manages each step with greater visibility but tighter coupling. Supporting patterns such as the Transactional Outbox prevent message loss, while idempotency ensures repeated operations do not cause unintended side effects. Together, these techniques allow distributed systems to achieve eventual consistency without relying on a single coordinating database.

0 comments Read more at DEV Community

ProgrammingHacker News ·

Petals lets users run large AI models collaboratively via distributed network

Petals is a new platform that enables users to run large language models at home using a distributed, BitTorrent-style approach. Rather than requiring a single powerful machine, the system splits model layers across multiple volunteer computers on a network. This makes running state-of-the-art AI models accessible without expensive hardware. The project was shared on Hacker News, where it attracted community attention and discussion. More details are available at petals.dev.

0 comments Read more at Hacker News

ProgrammingDEV Community ·

Why Your AI Agent's Eval Set Matters More Than Its Prompt

Many AI agents ship with weak evaluation sets built from imagination rather than real-world failures, leaving developers blind to production issues. When an agent fails in deployment, the fix is incomplete until that specific failure is captured as a permanent test case. Effective eval sets should reflect actual traffic distribution, not just edge cases, and assert on meaningful behavioral properties rather than exact output strings. Because AI agents are non-deterministic, overly rigid tests tend to be ignored over time, which is worse than having no tests at all. Unlike prompts that can be copied or rewritten, a well-curated eval set encoding years of observed failures is a durable, competitive asset that survives every model swap or system change.

0 comments Read more at DEV Community