AI / Core OpenAI Codex Application Fundamentals Interview Questions
What are the built-in tools available in the OpenAI Responses API?
The Responses API ships with several built-in tools that the model can invoke automatically without you writing wrapper code. These tools connect the model to the real world and the developer's environment.
| Tool | What it does | Key use case |
|---|---|---|
| web_search | Fetches real-time, cited information from the internet | Research agents, shopping assistants, live data queries |
| file_search | Retrieves relevant content from uploaded document repositories with metadata filtering | RAG-style Q&A over documentation, contracts, PDFs |
| computer_use | Lets the model interact with a computer - click, type, navigate UI | Browser automation, GUI testing, UI-driven workflows |
| code_interpreter | Executes Python code in a sandboxed container; can produce charts/files | Data analysis, calculations, generating visualisations |
| remote MCP servers | Connects to any Model Context Protocol server over the internet | Custom tooling, enterprise API integrations, specialised data sources |
# Using web_search in the Responses API: from openai import OpenAI client = OpenAI() response = client.responses.create( model="gpt-5.5", tools=[{"type": "web_search_preview"}], input="What are the latest changes to Python's asyncio in 3.12?", ) print(response.output_text) # Using multiple tools together: response = client.responses.create( model="gpt-5.5", tools=[ {"type": "web_search_preview"}, {"type": "file_search", "vector_store_ids": ["vs_abc123"]} ], input="Summarise our internal Q3 report and compare it with industry trends.", )
Tools can be combined in a single request. The model decides when and how to invoke them - the developer does not need to manually manage tool invocation loops as with function calling in Chat Completions.
More Related questions...