---
title: "Wokey (@wokey) | Hashnode"
description: "Low-cost Claude & GPT API with offline-verifiable responses. Read the latest articles by Wokey on Hashnode."
url: "https://hashnode.com/@wokey"
---

@wokey

# Wokey

@wokey · Joined October 2026

Low-cost Claude & GPT API with offline-verifiable responses.

[Website](https://wokey.ai/)

### [Claude Code and the Anthropic SDKs Behind a Custom Base URL: Setup, Checks, and Gotchas](https://wokey.hashnode.dev/claude-code-and-the-anthropic-sdks-behind-a-custom-base-url-setup-checks-and-gotchas?utm_source=hashnode&utm_medium=feed)

ANTHROPIC\_BASE\_URL, credential headers, a no-key echo server, and why the model field is not proof

10 min read · [wokey.hashnode.dev](https://wokey.hashnode.dev/claude-code-and-the-anthropic-sdks-behind-a-custom-base-url-setup-checks-and-gotchas?utm_source=hashnode&utm_medium=feed)

[1h ago](https://hashnode.com/posts/claude-code-and-the-anthropic-sdks-behind-a-custom-base-url-setup-checks-and-gotchas/6af405f29868be71ba1a9b7f)

### [What a Coding-Agent Session Really Costs: Estimating Spend and Cache Hits from a /models Pricing Feed](https://wokey.hashnode.dev/what-a-coding-agent-session-really-costs-estimating-spend-and-cache-hits-from-a-models-pricing-feed?utm_source=hashnode&utm_medium=feed)

Coding agents like Claude Code, Codex CLI, and opencode don't send "a prompt". Every turn re-sends the system prompt, the tool schemas, and the whole conversation so far. A 40-turn session can easily

8 min read · [wokey.hashnode.dev](https://wokey.hashnode.dev/what-a-coding-agent-session-really-costs-estimating-spend-and-cache-hits-from-a-models-pricing-feed?utm_source=hashnode&utm_medium=feed)

[Nov 4](https://hashnode.com/posts/what-a-coding-agent-session-really-costs-estimating-spend-and-cache-hits-from-a-models-pricing-feed/6aeacbd62d48859dc446833b)

### [Verify LLM API Responses Offline: Checking an AWS Nitro Proof on Your Own Machine](https://wokey.hashnode.dev/verify-llm-api-responses-offline-checking-an-aws-nitro-proof-on-your-own-machine?utm_source=hashnode&utm_medium=feed)

A hands-on guide to capturing an LLM response together with its enclave-signed proof, checking it locally, and seeing exactly what that proof covers and what it leaves out.

10 min read · [wokey.hashnode.dev](https://wokey.hashnode.dev/verify-llm-api-responses-offline-checking-an-aws-nitro-proof-on-your-own-machine?utm_source=hashnode&utm_medium=feed)

[Oct 23](https://hashnode.com/posts/verify-llm-api-responses-offline-checking-an-aws-nitro-proof-on-your-own-machine/6adaf878e956456802667591)