The latest articles on DEV Community by Adrian Issac (@adrian_issac_7bb3d453abc6). https://dev.to/adrian_issac_7bb3d453abc6 https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3392285%2Ff75c0b80-c73c-4d60-a43b-a798c0181acd.jpg https://dev.to/adrian_issac_7bb3d453abc6 en Adrian Issac Sun, 25 Oct 2026 18:04:44 +0000 https://dev.to/adrian_issac_7bb3d453abc6/i-turned-a-claude-code-session-into-a-local-anthropic-compatible-api-5f61 https://dev.to/adrian_issac_7bb3d453abc6/i-turned-a-claude-code-session-into-a-local-anthropic-compatible-api-5f61 <p>I wanted tools like LiteLLM and Hermes Agent to talk to my Claude Code session as if it were a normal Anthropic API. So I built ccapi.</p> <p><a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fdaw6rlnrndg51kzxj6xe.gif" class="article-body-image-wrapper"><img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fdaw6rlnrndg51kzxj6xe.gif" alt=" " width="800" height="306"></a></p> <h2> What it does </h2> <p>ccapi exposes <code>/v1/messages</code> on localhost. Any client that speaks the Anthropic Messages API (the Anthropic SDK, LiteLLM, Hermes Agent) can point at it. Streaming, tool calls, images, PDFs and thinking come through.</p> <h2> How it works </h2> <ul> <li>Two Claude Code mods plus a small Bun server.</li> <li>Each incoming request runs as a Claude Code subagent.</li> <li>Client tools are registered as native Claude Code tools. When the client sends back a tool result, it resumes the same subagent.</li> <li>A systemd user unit keeps the tmux session alive for 24/7 use.</li> </ul> <h2> What I learned </h2> <ul> <li>The Claude Code mods API is early access, so anything can change.</li> <li>Keeping it up 24/7 needed a supervisor that restarts the session when it dies.</li> <li>Tool calls are the hard part: parked calls need deadlines and slot cleanup, or failures leak sessions.</li> </ul> <h2> Limits </h2> <ul> <li> <code>max_tokens</code> is approximate; <code>temperature</code>, <code>top_p</code> and <code>top_k</code> are ignored.</li> <li>If the server restarts, native continuation is lost and history is rendered as text.</li> <li>The system prompt always starts with the Agent SDK line.</li> </ul> <h2> Important caveat </h2> <p>This is a personal experiment, localhost only. Anthropic's terms do not allow routing third-party traffic through Free/Pro/Max plan credentials, so please do not expose or share it. ccapi never touches your OAuth token; it only uses Claude Code primitives (subagents, plugin tools, hooks).</p> <p>Repo: <a href="https://github.com/S-kkipie/ccapi" rel="noopener noreferrer">https://github.com/S-kkipie/ccapi</a></p> <p>Feedback welcome, especially on the subagent-per-request design.</p> ai opensource llm claude