Blog
Insights, updates and stories from our team
Tutorial2026/08/01
Codex CLI with a Custom Provider: config.toml Setup That Actually Works (2026)
Point Codex CLI at any OpenAI-compatible endpoint with a six-line model_providers block — why OPENAI_BASE_URL stopped working, wire_api explained, verification and troubleshooting.
Engineering2026/08/01
OpenAI-Compatible APIs Explained: base_url, Endpoints, and the Gotchas Nobody Documents
What 'OpenAI-compatible' actually promises, how base_url works in every SDK, which endpoints matter, and how to test a provider before trusting it.
Tutorial2026/08/01
How to Send PDFs and Images to Claude and GPT via API (2026 Guide)
Working code for PDF and image input with the OpenAI and Anthropic APIs — base64 data URLs, real token costs per page, and when to send a PDF vs. a screenshot.
Guide2026/07/10
How to Cut Claude API Costs in 2026: 7 Levers That Actually Work
A practical guide to reducing your Claude API bill — prompt caching, batching, model tiering, and discounted gateways, with real per-token math.
Tutorial2026/07/09
Claude Code with a Custom API: How ANTHROPIC_BASE_URL Works (2026 Guide)
Point Claude Code at any Anthropic-compatible endpoint with two environment variables — setup, verification steps, gotchas, and troubleshooting.
Pricing2026/07/08
Claude vs GPT API Pricing in 2026: Full Cost Tables (and How to Pay Less)
Official per-token prices for Claude Fable, Opus, Sonnet, Haiku and GPT-5.4/5.5/5.6 side by side — plus worked examples and discounted access.