How to Generate TypeScript Types from JSON — and Fix What Auto-Tools Miss

·2 views

Developers can quickly convert JSON API responses into TypeScript interfaces using browser-based tools like jsonviewertool.com, which nest objects into separate interfaces automatically. However, these generators only analyse the single sample provided, leading to several common inaccuracies in the output. Key failure points include nullable fields being inferred as just null, empty arrays typed as any[], missing optional fields, string literals typed broadly as string, and numeric IDs or enums left as plain numbers. For more reliable results, developers working on maintained codebases are advised to generate types from a JSON Schema or OpenAPI spec using dedicated tools such as json-schema-to-typescript or openapi-typescript. Sample-based generation is best treated as a draft starting point, with each field reviewed manually before the types are used in production.

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)

Run Tokens Can Fix Flaky Email Tests in CI Pipelines

Email automation tests frequently fail in CI environments because they select the 'latest' email rather than verifying that a message belongs to the current test run. When multiple jobs execute simultaneously, shared or staging inboxes become noisy, causing tests to open the wrong email and produce unreliable results. A developer on DEV Community recommends generating a unique run token at the start of each test and embedding it in the outgoing email so assertions can confirm message identity before extracting any links or codes. This approach outperforms timestamp-based windows, which can become ambiguous under heavy CI load, and makes failure logs easier to interpret. The pattern works regardless of whether teams use disposable addresses, shared staging inboxes, or purpose-built mail helpers.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer Builds AI Tool to Verify Carbon Credit Claims Using Satellite Data

A developer has created CarbonSense, an AI-powered platform designed to improve transparency in carbon project monitoring. The tool combines satellite imagery with Google's Gemini API to help users analyze environmental changes and understand complex climate data without requiring specialist expertise. Built using Next.js, React, and TypeScript, the platform translates raw environmental information into plain-language insights. A key design challenge was calibrating how confidently the AI should present findings, given the complexity and sensitivity of environmental data. The project is publicly available as a live demo and open-source repository on GitHub.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer Builds Free Browser-Based JSON-LD Generator to Automate Structured Data

A developer working on a technical SEO website grew frustrated with repeatedly writing JSON-LD structured data markup by hand for common Schema.org types. The repetitive workflow — choosing a schema type, writing the markup, validating it, and fixing errors — proved unscalable over time. Unable to find a lightweight tool that required no account or setup and worked instantly in the browser, the developer built one tailored to their own publishing needs. The free tool, available at seogeo.tech, covers the most common Schema.org types and generates clean JSON-LD output. The developer continues to refine it based on real-world usage and is seeking feedback from other developers and technical SEO professionals.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How Send16 rebuilt its email API to work autonomously with AI agents

Send16, an email platform, spent several weeks redesigning its API to be fully usable by AI agents without requiring human intervention at any step. The key obstacle was domain verification — SPF, DKIM, and DMARC setup — which agents cannot perform, causing every initial send attempt to fail. To fix this, the team introduced a sandbox sender using a shared domain that delivers emails only to the account owner, giving agents a real, abuse-proof first success with no DNS configuration needed. They also added a /api/me endpoint so agents can immediately identify their workspace, quota, and sandbox recipient, plus a plain-text llms.txt file documenting the raw HTTP contract for models that cannot use an SDK. For MCP clients like Claude Desktop and Cursor, Send16 ships a hosted tool server with lazy API key validation and per-request authentication threading to safely handle multiple users.

0 comments Read more at DEV Community