Perplexica hosting

An open answer engine — searches the web, then explains with citations.

Search & web agentsMITfrom $11.99/mo

In short

Perplexica hosting at AIHostBox means a dedicated, single-tenant Perplexica instance on our infrastructure: we install it, pin it to a reviewed release, put it behind TLS on your domain, back it up nightly and keep it patched. Plans start at $11.99 per month and the admin account, the data and any API keys are yours.

Cited answersAnswer modesSearXNG backendLocal models

Perplexica is an open-source answer engine by developer ItzCrazyKns, and in March 2026 the project was renamed Vane to avoid confusion with Perplexity; the code, features and MIT licence carried over, and the old image name still works. You ask a question, it queries SearXNG, reads the results and writes an answer with numbered citations. Speed, Balanced and Quality modes trade response time against how deeply it researches, and sources can be narrowed to the web, discussions or academic papers.

It also searches images and videos, accepts uploaded PDFs, text files and images to question alongside web results, and keeps a local history of past searches. The value of self-hosting is that the whole loop is yours: the metasearch queries, the pages read and the model calls. That only works if it sits somewhere always on: it is a web app your team opens from any browser, and its history and uploads need to live in one place rather than on whichever laptop ran the search.

Perplexica hosting plans

PlusRecommended minimum

4 GB RAM · 2 vCPU · NVMe

$11.99/mo

Order

ProTeams & heavier load

8 GB RAM · 3 vCPU · NVMe

$18.99/mo

Order

Every plan is a dedicated single-tenant instance with NVMe storage, nightly off-box backups, TLS and managed patching. Quarterly, semi-annual and annual terms are discounted at checkout. Model tokens are not included — you bring your own key.

How we run Perplexica

The itzcrazykns1337/vane container on port 3000 with SearXNG bundled inside, and its settings, uploads and search history on a persistent /home/vane/data volume.

What we configure for you

Model provider keys (OpenAI, Anthropic, Gemini, Groq) or an Ollama URL entered in the setup screen
Chat and embedding model defaults chosen for your provider
Slim image with SEARXNG_API_URL if you prefer to use a separate SearXNG plan
Persistent data volume so history and uploads survive upgrades
TLS on your subdomain

Which plan fits

Plus (4 GB) runs Vane and its bundled SearXNG for one person or a small team using hosted models. Pro (8 GB) gives headroom for several people researching at once, file uploads with larger documents, and Quality mode, which reads more pages per question.

The honest limits

No built-in accounts, so access control has to be agreed at deployment. Search results depend on the upstream engines SearXNG queries, which can rate-limit or block a server IP during heavy use.

What people run Perplexica for

Private research assistant

Answers with sources, without a subscription.

Complete the stack

SearXNG for search, Ollama for the model, Perplexica in front.

Academic source hunting

Restrict sources to papers and get a summary with links to each one.

Search inside your own apps

Call its search API from a script or bot and get an answer plus sources back as JSON.

Licence and version policy

Item	Position
Upstream licence	MIT
Sold as managed hosting	Yes
Tenancy	Single tenant — one customer per instance
Patching	Critical advisories within 24 hours, high severity within 72

Our full position on licences, including the applications we decline to sell, is on the licensing page.

Works well with Perplexica

More Search & web agents we host

Questions

Perplexica hosting: your questions

Is Perplexica the same thing as Vane?

Yes. The GitHub project was renamed from Perplexica to Vane in March 2026, and the Docker image moved from itzcrazykns1337/perplexica to itzcrazykns1337/vane. Functionality is unchanged and both image names continue to work. We deploy the current Vane release; the page keeps the Perplexica name because that is what most people still search for.

Which model providers can it use?

Vane works with local models through Ollama and with hosted providers including OpenAI, Anthropic, Google Gemini and Groq. You enter keys in its setup screen on first launch, and they are stored in your instance’s data volume. Answer quality depends heavily on the model; small local models summarise sources less reliably.

Does it need its own SearXNG?

The standard image bundles SearXNG inside the container, which is what we run by default. If you already have a SearXNG plan with us we can use the slim image instead and point SEARXNG_API_URL at it; that instance needs the JSON output format and the Wolfram Alpha engine enabled, which we set for you.

Can people log in, or is it open?

Vane has no user accounts yet; authentication is listed on the project’s roadmap. Anyone who can reach the page can use it and spend your model credits, so the instance should not be left publicly reachable. Talk to us about how you want access limited before you share the address.

Is there an API?

Yes. POST /api/search takes a query, the sources to use (web, academic or discussions), a chat and embedding model and an optional speed, balanced or quality mode, and returns the answer with its sources as JSON or a stream. It has no built-in authentication either, so treat the endpoint as private.

Run Perplexica without running a server

A dedicated Perplexica instance, deployed and maintained by us. Month to month, cancel whenever.

Browse the catalogue Talk to a human