If you want to Build LLM API with FastAPI and OpenAI<\/strong>, this guide will walk you through it step by step. Instead of another basic demo, we\u2019ll create a production-ready API that clients can actually use \u2014 complete with documentation, error handling, JSON requests, and proper structure.<\/p>\n By the end, you\u2019ll have a professional backend AI service you can showcase in your portfolio or offer to consulting clients.<\/p>\n In this guide, we\u2019re going to build a professional LLM API<\/strong> with FastAPI and OpenAI<\/strong>, step by step. By the end, you\u2019ll have:<\/p>\n An async FastAPI endpoint<\/strong> capable of handling multiple requests<\/p>\n<\/li>\n JSON and query-based inputs<\/strong> so your API is flexible<\/p>\n<\/li>\n Caching and health checks<\/strong> to save time, cost, and headaches<\/p>\n<\/li>\n Ready-to-use Python and JavaScript client examples<\/strong><\/p>\n<\/li>\n A smooth VSCode workflow<\/strong> with standalone testing tools<\/p>\n<\/li>\n<\/ul>\n Let\u2019s get started!<\/p>\n <\/p>\n Create a Python virtual environment and install your dependencies. This ensures your project is isolated and professional<\/strong>.<\/p>\n This class wraps the OpenAI API so your endpoints stay clean and reusable.<\/p>\n Here we combine async endpoints<\/strong>, caching<\/strong>, health checks<\/strong>, and flexible input<\/strong> handling.<\/p>\n Recommended VSCode Extensions:<\/strong><\/p>\n Python<\/strong> \u2013 for running FastAPI and IntelliSense<\/p>\n<\/li>\n REST Client<\/strong> \u2013 send Thunder Client<\/strong> \u2013 lightweight Postman alternative<\/p>\n<\/li>\n Pylance<\/strong> \u2013 type checking and autocompletion<\/p>\n<\/li>\n<\/ul>\n Example ###<\/span><\/p>\n POST http:\/\/127.0.0.1:8000\/chat { Use caching<\/strong> for repeated prompts to reduce API costs.<\/p>\n<\/li>\n Combine GET\/POST<\/strong> endpoints for flexible client usage.<\/p>\n<\/li>\n Always include health checks<\/strong> in production APIs.<\/p>\n<\/li>\n Use FastAPI docs<\/strong> ( Share <\/p>\n Once you build LLM API with FastAPI and OpenAI, the next professional move is backing it up to GitHub. This protects your work, creates version history, and makes it easier to share your API with clients or collaborators.<\/p>\n Create a copy and paste the files to the gitignore file<\/><\/p>\n Never upload your <\/p>\n Create repo on GitHub (leave it empty).<\/p>\n Now your FastAPI and OpenAI LLM API project is safely backed up and ready to showcase in your portfolio or share with clients.<\/p>\n <\/p>\n When you build LLM API with FastAPI and OpenAI, relying on official documentation ensures your implementation follows best practices, stays secure, and remains production-ready. Be sure to reference the With this setup, you have a fully professional LLM API<\/strong>:<\/p>\n Async, cached, and ready for multiple clients<\/p>\n<\/li>\n Flexible input methods (JSON POST + GET query)<\/p>\n<\/li>\n Health check endpoint for monitoring<\/p>\n<\/li>\n Interactive documentation and client-ready examples<\/p>\n<\/li>\n<\/ul>\n You\u2019re ready to ship your own AI API<\/strong> or let clients plug in immediately!<\/p>\n","protected":false},"excerpt":{"rendered":" Why Build LLM API with FastAPI and OpenAI? If you want to Build LLM API with FastAPI and OpenAI, this guide will walk you through […]<\/p>\n","protected":false},"author":1,"featured_media":1407,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[9,13,28,27],"tags":[20,22,25,12,23,26,24],"class_list":["post-1405","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai","category-instructions","category-llm","category-python","tag-agentic-ai","tag-ai","tag-api","tag-instructions","tag-llm","tag-openai","tag-python"],"yoast_head":"\nProject Structure for a Production LLM API<\/h2>\n
\n
\nStep 1: Install Dependencies to Build LLM API with FastAPI and OpenAI<\/h2>\n
<?php\n\n# Create and activate venv\npython -m venv venv\nsource venv\/bin\/activate # macOS\/Linux\n# venv\\Scripts\\activate # Windows\n\n# Install dependencies\npip install fastapi uvicorn openai python-dotenv\n<\/code><\/pre><\/p>\nStep 2: Create the LLM Client (OpenAI Integration) (
llm.py<\/code>)<\/h2>\n<?php\n\nimport os\nimport logging\nfrom openai import OpenAI, OpenAIError\nfrom dotenv import load_dotenv\n\nload_dotenv()\nlogger = logging.getLogger(__name__)\nlogger.setLevel(logging.INFO)\n\ndef get_api_key() -> str:\n api_key = os.getenv("OPENAI_API_KEY")\n if not api_key:\n logger.error("OPENAI_API_KEY not set")\n raise ValueError("OPENAI_API_KEY environment variable not set")\n return api_key\n\nclass LLMClient:\n def __init__(self, api_key: str, model: str = "gpt-4o-mini"):\n self.model = model\n try:\n self.client = OpenAI(api_key=api_key)\n except OpenAIError as e:\n logger.exception("Failed to initialize OpenAI client")\n raise e\n\n def chat(self, prompt: str) -> str:\n try:\n response = self.client.chat.completions.create(\n model=self.model,\n messages=[{"role": "user", "content": prompt}]\n )\n return response.choices[0].message['content']\n except Exception as e:\n logger.exception("Error generating response")\n return f"Error generating response: {e}"\n<\/code><\/pre><\/p>\nStep 3: Build LLM API with FastAPI and OpenAI \u2014 The Main App (
main.py<\/code>)<\/h2>\n<?php\n\nfrom fastapi import FastAPI, HTTPException, Query, Request\nfrom pydantic import BaseModel\nimport logging\nfrom app.llm import LLMClient, get_api_key\nfrom openai import OpenAIError\nimport asyncio\nfrom functools import lru_cache\n\napp = FastAPI(title="Pro LLM API", version="1.0")\nlogger = logging.getLogger("api")\nlogger.setLevel(logging.INFO)\n\n# Models\nclass ChatRequest(BaseModel):\n prompt: str\n\nclass ChatResponse(BaseModel):\n response: str\n\n# Initialize LLM client\ntry:\n api_key = get_api_key()\n llm_client = LLMClient(api_key)\nexcept Exception as e:\n logger.exception("Failed to initialize LLM client")\n llm_client = None\n\n# Optional caching\n@lru_cache(maxsize=128)\ndef cached_chat(prompt: str) -> str:\n return llm_client.chat(prompt)\n\n# Health check\n@app.get("\/health")\nasync def health_check():\n status = "ok" if llm_client else "LLM client unavailable"\n return {"status": status}\n\n# Unified chat endpoint (GET + POST)\n@app.api_route("\/chat", methods=["GET", "POST"], response_model=ChatResponse)\nasync def chat_endpoint(request: Request, prompt: str = Query(None)):\n if not llm_client:\n raise HTTPException(status_code=500, detail="LLM client not available")\n\n # Determine prompt\n if request.method == "POST":\n data = await request.json()\n prompt_val = data.get("prompt")\n if not prompt_val:\n raise HTTPException(status_code=422, detail="POST request must include 'prompt'")\n else: # GET\n if not prompt:\n raise HTTPException(status_code=422, detail="GET request must include 'prompt'")\n prompt_val = prompt\n\n # Run in executor for async\n loop = asyncio.get_event_loop()\n try:\n llm_response = await loop.run_in_executor(None, cached_chat, prompt_val)\n return ChatResponse(response=llm_response)\n except OpenAIError as e:\n logger.error(f"OpenAI API error: {e}")\n raise HTTPException(status_code=502, detail="OpenAI API error")\n except Exception as e:\n logger.exception(f"Unexpected error: {e}")\n raise HTTPException(status_code=500, detail="Internal server error")\n<\/code><\/pre><\/p>\nStep 4: How Clients Use Your LLM API<\/h2>\n
Python<\/h3>\n
<?php\n\nimport requests\n\nurl = "http:\/\/127.0.0.1:8000\/chat"\nresp = requests.post(url, json={"prompt": "Tell me a joke"})\nprint(resp.json()["response"])<\/code><\/pre>\n<\/div>\n<\/div>\nJavaScript<\/h3>\n
<span class="hljs-title function_">fetch<\/span>(<span class="hljs-string">"http:\/\/127.0.0.1:8000\/chat"<\/span>, {
\n<span class="hljs-attr">method<\/span>: <span class="hljs-string">"POST"<\/span>,
\n<span class="hljs-attr">headers<\/span>: {<span class="hljs-string">"Content-Type"<\/span>: <span class="hljs-string">"application\/json"<\/span>},
\n<span class="hljs-attr">body<\/span>: <span class="hljs-title class_">JSON<\/span>.<span class="hljs-title function_">stringify<\/span>({ <span class="hljs-attr">prompt<\/span>: <span class="hljs-string">"Give me a motivational quote"<\/span> })
\n})
\n.<span class="hljs-title function_">then<\/span>(<span class="hljs-function"><span class="hljs-params">res<\/span><\/span> => res.<span class="hljs-title function_">json<\/span>())
\n.<span class="hljs-title function_">then<\/span>(<span class="hljs-function"><span class="hljs-params">data<\/span><\/span> => <span class="hljs-variable language_">console<\/span>.<span class="hljs-title function_">log<\/span>(data.<span class="hljs-property">response<\/span>));
\n<\/code><\/div>\n<\/div>\n
\nStep 5: Testing the FastAPI LLM API in VSCode<\/h2>\n
\n
.http<\/code> requests directly from VSCode<\/p>\n<\/li>\n.http<\/code> File for Quick Testing:<\/strong><\/p>\nGET http:\/\/127.0.0.1:8000\/chat?prompt=Hello<\/code><\/p>\n
\nContent-Type: application\/json<\/p>\n
\n“prompt”<\/span>: “Tell me a joke”<\/span>
\n}<\/p>\n<\/div>\n<\/div>\n
\nStep 6: Deploying Your LLM API to Production<\/h2>\n
\n
\/docs<\/code>) for interactive client testing.<\/p>\n<\/li>\n.http<\/code> or Postman collections for plug-and-play API testing.<\/p>\n<\/li>\n<\/ul>\nStep 7: Back Up Your Build LLM API with FastAPI and OpenAI to GitHub Using VSCode<\/h2>\n
1\ufe0f\u20e3 Create
.gitignore<\/code> FIRST<\/h3>\n.gitignore<\/code> file in your project root:<\/p>\nvenv\/<br\/>
\n.<span class="hljs-built_in">env<\/span><br\/>
\n__pycache__\/<br\/>
\n*.pyc<br\/>
\n<\/code><\/div>\n<\/div>\n.env<\/code> file or virtual environment. Your OpenAI API key should always stay private.<\/p>\n
\n2\ufe0f\u20e3 Initialize Git<\/h3>\n
git <span class="hljs-keyword">init<\/span>
\n<\/code><\/div>\n<\/div>\n
\n3\ufe0f\u20e3 Add and Commit<\/h3>\n
git <span class="hljs-keyword">add<\/span> .
\ngit <span class="hljs-keyword">commit<\/span> <span class="hljs-operator">-<\/span>m "Initial commit - FastAPI LLM API project"
\n<\/code><\/div>\n<\/div>\n
\n4\ufe0f\u20e3 Create GitHub Repository<\/h3>\n
\n5\ufe0f\u20e3 Connect and Push<\/h3>\n
<span class="hljs-selector-tag">git<\/span> <span class="hljs-selector-tag">remote<\/span> <span class="hljs-selector-tag">add<\/span> <span class="hljs-selector-tag">origin<\/span> <span class="hljs-selector-tag">https<\/span>:<span class="hljs-comment">\/\/github.com\/YOUR_USERNAME\/YOUR_REPO_NAME.git<\/span>
\n<span class="hljs-selector-tag">git<\/span> <span class="hljs-selector-tag">branch<\/span> <span class="hljs-selector-tag">-M<\/span> <span class="hljs-selector-tag">main<\/span>
\n<span class="hljs-selector-tag">git<\/span> <span class="hljs-selector-tag">push<\/span> <span class="hljs-selector-tag">-u<\/span> <span class="hljs-selector-tag">origin<\/span> <span class="hljs-selector-tag">main<\/span><\/code><\/div>\n<\/div>\n
\nOfficial Documentation to Build LLM API with FastAPI and OpenAI<\/h2>\n
\nFastAPI documentation<\/a>
\nfor framework guidance, the
\nOpenAI API documentation<\/a>
\nfor model integration, the
\nPydantic documentation<\/a>
\nfor request validation, and the
\nUvicorn ASGI server documentation<\/a>
\nfor running your application in development and production environments.<\/p>\nIf you\u2019re new to backend development, read my guide on Building REST APIs with Python<\/strong><\/a>.<\/h3>\n
\n\u2705 Conclusion<\/h3>\n
\n