Morphic
Morphic is an open-source AI search engine under Apache-2.0 that answers with grounded, cited results and renders them as a generative interface with images, grids and headings rather than plain text. It uses OpenAI, Anthropic, Google, Ollama or any OpenAI-compatible model, and Tavily, SearXNG, Brave or Exa for search. Docker Compose starts it with SearXNG, PostgreSQL and Redis, and it supports sign-in, chat history and shareable results.
What does Morphic do?
Morphic is "an AI-powered search engine with a generative UI." It gives "AI-powered search with grounded, cited answers", and its answers render "rich inline components (source-credited images, grids, headings) live from a streamed JSON spec, beyond plain markdown." It has quick and adaptive search modes, and a model selector that detects available providers: OpenAI, Anthropic, Google, Ollama, Vercel AI Gateway and OpenAI-compatible services.
Search runs through Tavily, SearXNG, Brave or Exa. Chat history is stored in PostgreSQL, results can be shared at unique URLs, files can be uploaded, and users sign in through Supabase Auth, with a guest mode for anonymous use. The README points to a privacy document that says what a self-hosted instance sends out and what is off by default.
Key facts
- Licence: Apache-2.0, with no added conditions.
- Grounded, cited answers rendered as a generative interface of images, grids and headings.
- Quick and adaptive search modes.
- Models: OpenAI, Anthropic, Google, Ollama, Vercel AI Gateway and OpenAI-compatible providers, detected automatically.
- Search providers: Tavily, SearXNG, Brave and Exa.
- Chat history in PostgreSQL; shareable result URLs; file upload; Supabase sign-in with a guest mode.
- Install: Docker Compose starts PostgreSQL, Redis, SearXNG and Morphic together, so no search API key is needed; at least one model key is required. Local development uses Bun.
- Latest tagged release when read: v1.7.0 on 2026-09-15; 9,125 stars on 2026-09-17.
What does it replace, and where does it fall short?
Morphic replaces Perplexity for someone who wants cited answers presented as a rich page rather than a block of text, with user accounts and shareable links, on their own deployment. It is the pick for sign-in and a choice of four search providers on the open-source alternatives to Perplexity page.
Where it falls short: chat history, authentication and file upload need extra configuration beyond the quick start, per the README. It needs a database and Redis, so it is heavier to run than Vane. No mobile app or agent features are described.
How does Morphic run?
Docker Compose is recommended: it starts PostgreSQL, Redis, SearXNG and Morphic, so no search API key is needed, but you must set at least one model provider key such as OPENAI_API_KEY. Local development uses Bun with an OpenAI key and a Tavily key. A one-click deploy button targets a hosting company. The README describes no paid tier from the maintainer.
Who is Morphic for?
A team that wants a private Perplexity with accounts, history and shareable answers, and is happy to run Postgres and Redis. One person on a laptop may find Vane simpler.
What limits does the README state?
None stated as limitations. The README notes that chat history, authentication, file upload and other features need configuration described in a separate document.
Questions people ask
Is Morphic open source?
Yes. Morphic is Apache-2.0 with no added conditions, so you may run, change and redistribute it. The README describes no paid tier from the maintainer; a hosted demo address appears in the repository's metadata rather than in the README itself.
Does Morphic need a search API key?
Not with Docker Compose, which starts a SearXNG instance alongside Morphic, so the README says no additional search API key is needed. Tavily, Brave and Exa are optional providers. You do need at least one model provider key.
Can Morphic use local models?
Yes. Ollama is in its list of model providers, alongside OpenAI, Anthropic, Google, Vercel AI Gateway and any OpenAI-compatible service, and the model selector detects which of them are available. Pairing Ollama with the bundled SearXNG keeps both model and search on your own servers.
How does Morphic compare with Perplexity?
Morphic gives cited answers with a richer, generated layout on your own servers, with accounts and shareable links, using the model and search provider you pick. Perplexity is hosted with its own index and apps. The open-source alternatives to Perplexity page compares it with Vane and Khoj.
What is a generative UI in Morphic?
Instead of returning only Markdown text, Morphic streams a structured description of the answer and renders it as components: source-credited images, grids and headings, built live as the answer arrives. That is the README's own description of the feature.
Sources
- Morphic README and Apache-2.0 LICENSE: github.com/miurla/morphic, read 2026-09-17.
Compared with the others
On the open-source alternatives to Perplexity page, Morphic is the pick for a richer answer page with accounts. Apache-2.0, cited answers rendered as a generated layout of images, grids and headings, with Tavily, SearXNG, Brave or Exa, Supabase sign-in and shareable results.
Also on that page: Vane for the closest perplexity shape, self-hosted, Khoj for answers from your documents as well as the web.
More answer engines
An AGPL-3.0 personal AI that answers from the web and your own documents with any local or hosted model, builds custom agents, sends scheduled research by email, and runs self-hosted or in its maintainers' cloud.
An MIT answer engine, formerly Perplexica, that searches the web through a bundled SearXNG and answers with cited sources, using local models through Ollama or hosted ones, in speed, balanced and quality modes, on your own hardware.
Added September 17, 2026. Every claim above comes from the project's README, LICENSE or model card, read on September 17, 2026, or from the GitHub API on the date shown in the panel. Found an error? Write to reveneau@licheo.com and it is fixed in the next weekly pass. Repository: github.com/miurla/morphic.
Running one of these inside your own environment, with your own data and your own security rules, is the kind of work Reveneau does. Read how a forward deployed engagement works.