Skip to content
publik.
Browse appsAppsPricingSupport buildersSupport
+Publish a repoPublish
Browse appsHow it worksPublish a repoSupport buildersGet helpPricingDevelopersGitHubPrivacy
© 2026 Publik
← All apps

Astro

An answer engine that can actually use your browser — no key to paste

AI: local or publik API

Astro interface

On your phone?

You install Astro from a computer. Send yourself the link and open it there.

Before you start: A Mac with Apple silicon, or a 64-bit Windows PC — both download and run directly, no toolchain to install first. Windows is unsigned, so SmartScreen needs one click through (More info, then Run anyway). Astro's built-in agent defaults to publik API with no key to paste — linking your publik account gives $0.05 of free use once, then the app pays publik's published price per use, and your own key still works

Available for macOS and Windows.

◆Built on Publik APIBuild yours →

Runs on publik API

  • AI chat & tools: Chat completions with tool calls, JSON schemas and images, priced per tier in dollars.
  • No key to paste.
  • Linking your publik account gives five cents of free use, once.
  • Then Astro pays publik's published price per use, and your own key still works.
How pricing works →

Checked on a real device on September 28, 2026.

Vote on Astro
14
Read the install guide→Open in GitHub↗

Compared with

P

Perplexity Comet

Perplexity Pro; Max is $200/mo

$20/mo

What a month of AI costs

  • Perplexity Comet$20.00/mo
  • publik API$4.33/mo

publik API: $15.67 a month less than Perplexity Comet.

100 uses a week at the Balanced level · publik API is cheaper up to about 461 uses a week

Make this yours→Fork Astro, change it, publish your version. About 30 minutes. No experience needed.

597 downloads through publik

Having trouble? Tell us

Where Astro’s AI runs, and what it costs

Prices are for one typical use: one request of about 1,500 words sent and 375 words back. A higher quality score is better.

On your computer

Price
Free per use. Your computer does the work.
Runs with
Ollama, LM Studio
Memory
Llama 3.1 8B is a 4.9 GB download at 4-bit. gpt-oss 20B runs on systems with as little as 16GB memory. Faster with Apple silicon or a graphics card.
Quality
Llama 3.1 8B: 7

publik API

Price
$0.0100 per use, $10.00 per 1,000 usespublik-balanced, the default level
Quality
MiMo-V2.6-Pro: 46
Setup
Built in. No key to paste; linking your publik account gives $0.05 of free use, once.

Your own key

Price
$0.0013 per use, $1.31 per 1,000 usesMiMo-V2.6-Pro at OpenRouter’s list price, before its fee for buying usage
Quality
MiMo-V2.6-Pro: 46
Setup
Open a provider account, add a card, paste the key into Astro.

Quality and price, side by side

Quality score and cost per 1,000 typical uses for local models and the three publik API levels
ModelQualityPer 1,000 uses
On your computer (Ollama, 4-bit download size)
Granite 4.2 3B2.2 GB9$0
Phi-4 Mini2.5 GB6$0
Llama 3.1 8B4.9 GB7$0
gpt-oss 20B14 GB9$0
Gemma 4 31B20 GB19*$0
Qwen3.5 35B-A3B24 GB19*$0
publik API
publik-fastGLM-5.3 Flash42$1.00
publik-balancedMiMo-V2.6-Pro · Astro46$10.00
publik-smartGPT-6 Sol48$18.00

Quality: Artificial Analysis Intelligence Index v4.3.2, read 2026-09-22 (publik API models 2026-09-25); * = estimated by Artificial Analysis. Sizes: the Ollama library, read 2026-09-22.

What you pay for

  • On your computer: nothing per use. You pay in disk space, memory and electricity, at lower quality.
  • publik API: publik’s published price for each use, in dollars, from your publik balance. It is above the model’s cost; the difference runs publik and pays the app’s builder.
  • Your own key: the provider’s price, billed to an account you keep with the provider.
Install Astro: publik API is built in →How publik API pricing works →

How to install Astro

Every step written out. No terminal experience needed. Pick your setup.

  • How to install Astro on Mac →
  • How to install Astro on Windows →

README

Open in GitHub ↗
Astro

Astro

An answer engine that can actually use your browser

Perplexity-style search, deep research and multi-model comparison — on top of a browser its agent can genuinely drive. Bring your own key, or reuse a subscription you already pay for. Search costs nothing and never leaves your machine.

Setup · Attribution


What it is

Two open-source projects, joined at the seam where each is weakest.

Simplicity is a very good answer engine with no browser under it. BrowserOS is a real browser whose agent can drive pages, but it had no retrieval at all — its agent had to open Google and scrape the results page like a person.

Astro puts Simplicity's UI and search pipeline on top of BrowserOS's browser automation. You get an answer engine with citations and an agent that can open tabs, click, read and fill forms — in the same window.

SearchFree and private. A local SearXNG is provisioned on first use — no Docker, no account, no search API key
Deep ResearchPlans queries, searches in parallel, reads the best sources in full, refines, and cites
Model CouncilSeveral models answer from one shared retrieval; a chair compares them and surfaces the disagreements
Browser use16 CDP tools — tabs, navigation, clicking, reading, forms, screenshots
Your modelsClaude Code, Codex and Copilot subscriptions; any API key; or fully local via Ollama / LM Studio

Scraped page content is fenced as untrusted data before it reaches a model, so a web page cannot issue instructions to your agent.

What is not built

Stated plainly, because the navigation shows more than the code does:

  • Computer — Comet's version runs tasks in a cloud sandbox and keeps working with the laptop closed. Not built. The pieces that exist are deep_research, model_council and the browser tools, all synchronous and local.
  • Spaces, Artifacts, Skills, Memory — navigation entries only.
  • File generation (.pptx / .xlsx) — absent here and in both upstreams.

Licences

Astro is AGPL-3.0, inherited from BrowserOS, which is based on Chromium and uses patches from ungoogled-chromium.

The vendored Simplicity/Vane code is MIT, © ItzCrazyKns. ATTRIBUTION.md carries the notice, maps every ported file to its source, and documents where this fork deliberately diverges from both.