Morphic
Open SourceOpen-source AI answer engine with a generative UI — a Perplexity-style search experience built on Next.js and the Vercel AI SDK, with pluggable search providers (Tavily, SearXNG, Exa), model switching, and chat history.
Why we recommend Morphic
Morphic is one of the best open starting points for the 'AI search / answer engine' product shape that Perplexity made mainstream: ask a question, watch it search the web, and get a streamed, cited answer with follow-ups — implemented with the Vercel AI SDK's generative UI so the model's tool calls render as real interface components, not text. It abstracts the retrieval layer behind pluggable providers (Tavily, Exa, or self-hosted SearXNG for zero-API-cost search) and supports multiple LLM backends including Ollama for local models. Reading it teaches the RAG-over-live-web pattern; forking it gives you a working vertical search product to specialize.
Best use cases
Best for: developers building domain-specific answer engines (legal, docs, e-commerce search), teams prototyping RAG products who want retrieval + streaming + citation UI pre-assembled, and self-hosters who want a private Perplexity on SearXNG + Ollama.
Honest limitations
Not for: turnkey commercial deployment — no billing, quotas or abuse protection, and web-search API costs (or SearXNG ops) plus LLM costs are immediately yours. Answer quality is bounded by the search provider you plug in. Primarily one maintainer plus community; expect to own hard bugs yourself.
Are you the maker of Morphic? Claim this listing →
More in AI Apps
Vercel's official open-source AI chatbot template — Next.js App Router with the AI SDK, streaming chat UI, model-provider switching (OpenAI, Anthropic, xAI and more), auth, and chat history persistence.
Open-source AI chat interface by Mckay Wrigley — multi-model chat (OpenAI, Anthropic, local models via Ollama), folders and prompts, file attachments, and Supabase-backed persistence. One of the most-forked AI UI codebases on GitHub.
Open-source template for AI apps that generate and run code — a v0/Claude-Artifacts-style interface where LLM output executes in secure E2B sandboxes, with Next.js, the AI SDK, and support for multiple model providers.
Open-source TypeScript/React library and starter for AI chat interfaces — composable primitives in the shadcn style for messages, threads, streaming, tool calls and generative UI, with AI SDK and LangGraph integrations.
Official LangChain.js starter for Next.js — working examples of chat, structured output, agents, and retrieval-augmented generation with streaming, ready to deploy on Vercel. The canonical way to start a LangChain web app. Open-source.
Next.js AI boilerplate shipping multiple ready-made AI demo apps (chat, PDF, vision, audio, image generation) on one codebase — auth, payments, analytics and prompts included, aimed at launching niche AI wrappers fast. Paid, one-time purchase.