Morphic
AI-powered search engine with a generative UI that gives cited answers rendered as rich inline components.
Morphic answers with cited sources and renders images, grids and headings from a streamed spec.
Morphic is an AI search engine built on Next.js whose answers render as rich components rather than plain markdown. It suits developers who want to run or fork a Perplexity-style app with their choice of model and search provider. It is a web app you deploy, with a hosted instance linked from the repository.
What it is
You ask a question, Morphic searches through Tavily, SearXNG, Brave or Exa, and a model of your choice writes a cited answer. The answer can include source-credited images, grids and headings built from a streamed JSON spec. It stores chat history in PostgreSQL and lets you share a result by unique URL.
Who it's for
It suits developers and small teams who want a customisable AI search front end and can run Docker Compose or deploy to Vercel. It does not suit non-technical users who want a finished product with support, because setup involves environment files and several services.
Key features
Generative UI answers
Answers render inline images, grids and headings live from a streamed JSON spec.
Quick and Adaptive modes
Pick a faster search mode or an adaptive one for harder questions.
Model selector
Choose among OpenAI, Anthropic, Google, Ollama, Vercel AI Gateway and OpenAI-compatible providers.
Several search backends
Use Tavily, SearXNG, Brave or Exa as the search provider.
Shareable results
Share a search result through a unique URL, with chat history in PostgreSQL.
Accounts and guest mode
Built-in accounts through Better Auth or Supabase Auth, plus anonymous guest use.
Up and running in 3 steps
Our honest verdict
What's great
- The generative UI produces richer answers than plain text, including source-credited images.
- Docker Compose starts Morphic, PostgreSQL, Redis and SearXNG together, so no separate search API key is needed.
- A PRIVACY page documents what a self-hosted instance sends outbound and what is off by default.
- A one-click Vercel deploy button is provided if you prefer not to run servers.
Where it falls short
- You must supply at least one AI provider API key, so every answer costs model tokens unless you use a local Ollama model.
- The full stack runs four containers (app, PostgreSQL, Redis, SearXNG), which is real operating overhead for a personal tool.
- Features like chat history, authentication and file upload need extra configuration beyond the quick start.
- It is a developer-oriented codebase using Bun and Next.js, so customising it assumes TypeScript skills.
Pricing & license
Morphic is free under Apache-2.0 and the repository shows no paid plan; it links a hosted instance at https://chat.morphic.sh whose terms are not covered in the README. Running it yourself costs your servers plus any model and search API usage.
Open remote AI-training roles
870 open remote AI-training roles, updated November 14, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Can I use Morphic without paying for search?
Yes. Docker Compose bundles SearXNG, so no extra search API key is required, though you still need a model provider key or a local Ollama model.
Does Morphic need a database?
Chat history is stored in PostgreSQL, which the Docker Compose setup starts for you.
Is there a hosted version?
The repository links chat.morphic.sh as its website, and you can also deploy your own to Vercel.
