Claude API Developer — Build a Client-Facing AI Assistant

We're looking for a developer with hands-on Anthropic Claude API experience to build a working prototype of a client-facing AI assistant. The scope is deliberately small and well-defined — a single chat application, built cleanly, that we can extend afterwards. This is the first phase of a larger build, and we intend to continue with whoever delivers this well. 𝐖𝐡𝐚𝐭 𝐲𝐨𝐮'𝐥𝐥 𝐝𝐞𝐥𝐢𝐯𝐞𝐫: • A working chat interface backed by the Claude Messages API • Streaming responses (no waiting for the full reply) • Conversation history maintained across turns • A well-structured system prompt kept in config, easy for us to edit • At least one working tool call (function calling) wired to a simple external data source a public API or a JSON file is fine • Basic error handling: rate limits, timeouts, failed tool calls Source code in a private GitHub repo • A short README: setup, environment variables, how to swap the API key • A live demo link (Vercel, Render, Railway your choice) 𝐒𝐭𝐚𝐜𝐤: Your call. Next.js/React or Python/FastAPI both work for us. Keep it simple and readable no heavy framework layers we'd have to unpick later. 𝐖𝐡𝐚𝐭 𝐰𝐞 𝐩𝐫𝐨𝐯𝐢𝐝𝐞: API key, a one-page brief on the use case, and fast feedback. We're responsive during IST business hours and won't leave you waiting on answers. 𝐘𝐨𝐮'𝐫𝐞 𝐚 𝐠𝐨𝐨𝐝 𝐟𝐢𝐭 𝐢𝐟 𝐲𝐨𝐮: • Have shipped something real on the Claude or OpenAI API not just tutorials • Understa nd streaming, token limits, and tool/function calling in practice •Write code someone else can pick up and extend • Communicate clearly and flag blockers early 𝐓𝐨 𝐚𝐩𝐩𝐥𝐲, 𝐢𝐧𝐜𝐥𝐮𝐝𝐞: 1. A link to something you've shipped on the Claude or OpenAI API repo or live demo, not a description 2. Which stack you'd use here and why, in one or two sentences 3. Your realistic timeline Applications that open with a generic agency pitch and no shipped work will not be reviewed.