Search the web with sourced answers, visual results, article summaries and explanations, with a separate search API for developers.
Andi
Explore features, practical uses and pricing below.
Andi is an AI search assistant that combines web results with answers linked to their sources. Users can ask a question in ordinary language, search for a page or explore the material behind an answer. The interface brings search and reading together, making it useful for students, writers and professionals collecting background information from the open web.
Its consumer search interface and developer API serve different workflows. Someone researching a topic can use Andi directly in a browser. A developer can use the separate API to add web retrieval to an application or agent. Both belong to the same search product family, but their access and billing should be considered separately.
The official help guide describes natural-language questions with source links and ordinary web searches with result cards. Desktop users can choose Feed, Grid or List views. A site-specific search narrows the task to a named website, while the exact-query command offers a keyword-search route when that is more useful than an AI answer.
Result actions include Read, Summarize and Explain. Read provides a cleaner article view; Summarize produces a shorter account of a page; Explain aims to make complex material easier to understand. These actions support different stages of research: scanning a candidate source, reading the original context and clarifying unfamiliar material. Navigation commands can also take a user to a named website.
A writer preparing an introduction to urban heat could ask a specific question about the relationship between street trees and local temperature. They would inspect the linked sources, choose a relevant result and read it before using a summary to organize notes. An explanation can help unpack a technical term, while a site-specific query can focus further searching on an identified research institution.
The writer should record the source and publication date alongside each important claim. If two sources cover different cities, time periods or measurements, those differences belong in the notes rather than being flattened into one general conclusion. Andi helps locate and condense material; deciding whether a source supports a sentence remains part of the researcher's work.
The Andi Search API returns structured web results and offers a fetch endpoint for extracted page content. Developers can use source-attributed contextual output, automatic or selected search effort, and MCP tools. The API documentation covers authentication and integration. This route suits applications that need search results as inputs, rather than a person operating the consumer interface.
A source link makes an answer inspectable, but does not by itself establish that every sentence is complete or correct. Check original pages for qualifications and dates, especially where a summary omits context. For an application, evaluate whether the required pages can be retrieved and what happens when content cannot be extracted.
Andi describes consumer search as free. Developer access requires an API key and uses separate usage-based charges; consult the current offer before budgeting a workload. The privacy policy distinguishes anonymous search from optional account services and describes browser-local settings. Review those arrangements for the workflow being used, and keep developer credentials out of public client code.