A post shared on X is drawing attention to Perplexica, an open-source AI search engine presented as a full Perplexity AI clone that may even go further in some areas. Based strictly on the information included in the post, the project is positioned as a privacy-first alternative that runs entirely on the user’s own machine.
The central claim is simple: Perplexica offers the same kind of cited-answer experience and deep research workflow people associate with Perplexity, but without accounts, tracking, ads, or data collection. The post also says that no data leaves the user’s computer, making the software especially notable for users who care about local control and privacy.
What the post says Perplexica can do
🚨 Someone just open sourced a full Perplexity AI clone. And it might actually be better.
It's called Perplexica.
A privacy-first AI search engine that runs entirely on your machine. Same cited sources. Same deep research. Zero data leaving your computer.
You're paying… pic.twitter.com/d5d066V8pp
— Nav Toor (@heynavtoor) March 7, 2026
According to the tweet, Perplexica searches the entire web through SearxNG, a meta-search engine that can query Google, Bing, DuckDuckGo, and other sources at the same time. The idea is that the software reads top results, understands them, and then returns a cited answer with sources.
The post also lists several search and research modes. These include normal web search, academic papers, YouTube, Reddit, Wolfram Alpha, and a writing mode. That suggests the tool is not being framed as just another chatbot, but as a broader research interface designed for different types of information gathering.
Another feature highlighted in the tweet is support for PDFs, text files, and images. Users can upload those files and ask questions about them. The post also says Perplexica supports domain-specific search, which could be useful for people who already know which websites they want to search.
Beyond text-based research, the post claims that image and video search are built in. It also says that full search history is saved locally, reinforcing the privacy-focused pitch around the product.
Local-first, but not limited to one model stack
One of the more important details in the post is that Perplexica is said to work with Ollama for a 100% local setup, while also supporting OpenAI, Claude, Gemini, Groq, and any OpenAI-compatible API. That suggests the project is designed to be flexible rather than locked into a single model provider.
If that description is accurate, Perplexica would appeal to two very different groups at once: users who want a fully local setup, and users who prefer to connect the interface to external AI providers while still managing the search workflow themselves.
The install process highlighted in the post
The tweet emphasizes how simple setup is supposed to be. It says the software can be installed with a single command:
docker run -d -p 3000:3000 perplexicaAfter that, the instructions given in the post are to open a browser, go to localhost:3000, and use it as a private Perplexity-like experience running on the local machine.
A “Discover” feed and browser search replacement
The post also claims that Perplexica includes a Discover feed designed to surface interesting articles throughout the day. It describes that experience as a private, ad-free, AI-powered alternative to Google News.
Another notable claim is that the tool can be set as the default search engine in Chrome or Firefox. In other words, the post presents it not only as a research assistant, but as a possible replacement for conventional web search in day-to-day browsing.
The privacy angle is the main selling point
The most aggressive comparison in the post is the one made against Perplexity itself. The author contrasts Perplexity’s subscription price of $20 per month with Perplexica being described as free forever. The tweet also makes a strong privacy argument, saying that every search made on Perplexity is kept by that service, while every search made on Perplexica remains only with the user.
That framing is clearly meant to position Perplexica as an open, local, privacy-respecting alternative for people who like AI-assisted search but dislike sending their queries to a centralized platform.
Open-source traction mentioned in the post
The post says the project has already reached 27K GitHub stars, 2.8K forks, 764 commits, 44 contributors, and 31 releases. It also describes the project as actively maintained.
Finally, the tweet states that Perplexica is 100% open source and released under the MIT License.
What can be concluded from the post alone
Based only on the information shared in the tweet, Perplexica is being presented as a serious open-source attempt to recreate and possibly extend the Perplexity-style AI search experience. The biggest themes are clear: local execution, privacy, cited web answers, multi-mode research, and open-source flexibility.
That said, everything above reflects the claims made in the post itself. Without going beyond the tweet, the safest conclusion is that Perplexica is being promoted as a free, locally run, privacy-first AI search engine with web search, file analysis, multiple research modes, local history, and broad model compatibility.
