Only one slice of Google AI Pro can be rebuilt: the chat window. Over the Gemini API an agent builds that in a sitting, with file uploads, a model picker and a cost meter, and the Flash models have a free API tier, so light use can cost nothing. The other parts of the $19.99 aren't code: 5 TB of storage, Gemini inside Gmail and Docs, Flow video, NotebookLM, Jules and Antigravity limits, and perks such as YouTube Premium Lite. Judge the plan by which of those you use, because a self-hosted chat only replaces the first.
Before you buildTry the free Gemini app first, and Google AI Plus ($4.99) if you only want more room and storage. Pro is the plan to question if you use the storage and the Google apps rather than the chat.
Build me a private chat app to replace Google AI Pro. Requirements: - Node.js 22 + TypeScript: a Fastify server and a React UI built with Vite. `npm start` serves everything on http://127.0.0.1:4720. Single user, no accounts, no telemetry, local-first; listen on 127.0.0.1 only. - Two providers behind one interface: Gemini through the official `@google/genai` package (GEMINI_API_KEY and a comma-separated GEMINI_MODELS list in .env), and every model already pulled in Ollama on localhost:11434. Pick a model per chat. - Chat: streaming replies with Markdown and code highlighting (shiki), a stop button, regenerate, and edit-a-message to branch the chat from that point. When the API answers 429, show the retry time instead of failing silently. - Storage: better-sqlite3 in ./data/chats.db (chats, messages, model, input and output tokens); uploads in ./data/uploads/. Text goes as text; images, PDFs and audio under 15 MB go inline, larger files through the Gemini Files API. Full-text search over every chat with SQLite FTS5. - Projects: a folder of chats that share one system prompt and a few pinned files. - Cost meter: prices per model in ./prices.json, where free-tier models show $0.00 and a "free tier" badge. Show the cost of each reply, a monthly total, and refuse to send once MONTHLY_CAP_USD from .env is reached. - Import my Gemini export: given a Google Takeout ZIP, find the Gemini Apps activity file, print its structure and an entry count first, then add the chats it can reconstruct. Importing twice must not duplicate anything. - Out of scope: Gmail and Docs integration, video and music generation, voice, phone apps. - README: setup, where the data lives, and a plain note that on the free API tier Google may use prompts and responses to improve its products and humans may read them, while a billed Cloud project's prompts are not used.
Opens your agent with the prompt filled in, you press enter. Or copy it and paste it into an empty folder.
The first prompt builds the core. Run these one at a time in the same folder to grow it into your own Google AI Pro.
Add search grounding to the chat app in this folder. Requirements: - A per-chat "Search" toggle that enables Gemini's Google Search grounding tool in the `@google/genai` call; Ollama models don't get the toggle. - Show the sources the API returns as numbered links under the reply, and the search queries it ran. - Count grounded requests in the cost meter: the first 5,000 a month are free for Gemini 3 and newer, then $14 per 1,000. Put those numbers in ./prices.json, not in code. - Update the README with the link to Google's pricing page.
Add document notebooks to the chat app in this folder. Requirements: - A notebook is a named set of up to 50 files (PDF, Markdown, text) in ./data/notebooks/. - Chunk each file, embed the chunks with Ollama's `nomic-embed-text`, and store vectors in SQLite with `sqlite-vec`. - Each question retrieves the top 8 chunks; the reply cites them as [file, page] links that open the passage. - Re-index only files whose hash changed; show indexing progress. - Update the README.
Add a research mode to the chat app in this folder. Requirements: - A "Research" toggle asks for a budget in dollars (default 1.00), then loops: plan 3-6 sub-questions, run Google Search grounding for each, read the cited pages, then write the report. - Stop when the budget is spent and say so in the report. - Every claim links to a numbered source; list sources at the end. - Save the plan, the pages read and the cost with the chat so I can audit it. - Update the README.
Make the chat app in this folder usable from my phone. Requirements: - A responsive layout below 640px wide: chat list in a drawer, full-width input. - A web app manifest and icons so it installs to the home screen. - Reach it over Tailscale: the server keeps listening on 127.0.0.1 and the README explains `tailscale serve` to expose it to my devices only. - Never listen on 0.0.0.0 and add no login of its own. - Update the README with the Tailscale steps.
One $19.99 plan covers storage and AI together: terabytes for Drive, Gmail and Photos, Gemini inside the apps people already live in, Flow video and a YouTube Premium Lite subscription. For some people the storage alone justifies it, and leaving means moving years of mail and photos or paying for space elsewhere.
Estimate · Flash models on the free API tier and local Ollama models cost nothing beyond electricity. On the paid API, Gemini 4 Argon costs $2 in and $10 out per million tokens until Nov 30, 2026, then $4 and $20, so a typical reply is about a cent now and two cents after, roughly $9 or $18 a month at 30 messages a day. Gemini 3.8 Flash ($0.75 in, $3.75 out, rising on Jan 1, 2027) is far cheaper. The build session is token usage on your agent plan.
Checked 2026-10-19 against Google AI Pro's docs: 1, 2, 3, 4, 5, 6, 7
| Plan | Monthly | Annual, per month | What you get |
|---|---|---|---|
| Google AI Plus | $4.99 | — | 400 GB of storage, 2x the usage of non-subscribers, Deep Research, Gemini in Gmail, 200 Flow credits a month. |
| Google AI Pro | $19.99 | — | 5 TB of storage (Google's help page says 5 TB or 10 TB depending on the plan), 4x the usage of non-subscribers, Deep Research, 1,000 Flow credits and 10,000 Flow Music credits a month, extended Jules and Antigravity, $10 a month of Google Cloud credits, YouTube Premium Lite, Google Health Premium and Google Home Premium Standard. |
| Google AI Ultra | $99.99 | — | Up to 20x the usage of non-subscribers, 20 TB of storage on the lower tier and 30 TB on the higher, 10,000 or 25,000 Flow credits a month, $40 or $100 a month of Cloud credits, individual YouTube Premium. |
Anyone with a Google account can use the Gemini app for free with lower limits; Google lists Pro at 4x the usage of people without an AI plan.
Kinda. The core is a weekend build with the prompt on this page, but real gaps remain: Gemini inside Gmail, Docs, Vids and the other Google apps, 5 TB of storage shared by Drive, Gmail and Photos.
About $19.99 a month on the Google AI Pro plan, $239.88 a year (checked 2026-10-19).
Gemini inside Gmail, Docs, Vids and the other Google apps; 5 TB of storage shared by Drive, Gmail and Photos; Flow video and music generation, NotebookLM's higher limits, Jules and Antigravity; Deep Research as a one-click feature; The bundled perks: YouTube Premium Lite, Google Health Premium and Google Home Premium. If any of those matter to you, keep paying.
Yes: Open WebUI (free), LibreChat (open source), Jan (open source).