Show HN: Infinite โ€“ query your GA4/Stripe/PostHog data locally, on your machine Infinite OS, a self-hosted, local-first growth-analytics runtime, lets users query data from Google Analytics 4, Stripe, PostHog, and other sources in plain English on their own machine. The tool syncs data into a local Postgres database and uses the user's own Claude or Codex account for natural-language processing, ensuring data never leaves the machine. Maker River Khan released the open-source project on GitHub. Your growth data, on your own machine โ€” ask it anything in plain English. Infinite OS is a self-hosted, local-first growth-analytics runtime. It connects your data sources โ€” Google Analytics 4, PostHog, Stripe, Meta, Shopify, and read-only X public post metrics โ€” into a Postgres database on your own machine, keeps it synced, and lets you ask questions like "how many page views in the last 7 days?" or "which channels drove the most traffic?" and get real answers from governed, source-accurate metrics. ๐ŸŒ Site: infinite.fast https://infinite.fast ยท ๐• Project: @InfiniteOS https://x.com/InfiniteOS ยท Maker: @RiverKhan https://x.com/RiverKhan It is nota broad agent runtime, skill platform, or generic SQL tool. For natural-language questions it forwards your prompt to your own Claude or Codex account using your own credentials; it executes no tools or shell commands. Your data never leaves your machine unless you send it somewhere. Everything runs locally as one small stack: Postgres โ€” your synced growth data, sync state, jobs, schedules, and queryable views a rebuildable cache; the source of truth stays with the providers . app daemon โ€” a local HTTP API the CLI talks to. worker โ€” owns the scheduler and runs syncs as background jobs. CLI โ€” the operator shell + the chat interface. infinite encrypted connector credentials โ€” provider keys/tokens are encrypted at rest with your GROWTH OS ENCRYPTION KEY . You connect a source once, sync it, then ask questions. Answers come from a typed metric layer with authority/provenance rules โ€” not free-form SQL โ€” so a number is either source-accurate or honestly reported as unavailable. Once a source is connected and synced, just ask โ€” in plain English: "How many page views did I get in the last 7 days?" "What were my top pages this week?" "Which channels drove the most traffic last month?" "How many new users this week vs last week?" "What's my engagement rate over the last 30 days?" "How much revenue did I make in the last 30 days?" Stripe "Compare this week's traffic to the previous week." Infinite figures out the right metric, runs it against your synced data, and answers with the numbers + a short read โ€” citing the source and flagging anything it can't verify it won't guess . Install with one command macOS / Linux . It checks you have git, Node โ‰ฅ 20, and pnpm it won't install or change them for you , puts Infinite at ~/.infinite/app , drops an infinite command on your PATH, and runs setup: curl -fsSL https://raw.githubusercontent.com/Infinite-Labs-AI/infinite-os/main/scripts/install.sh | bash Then: infinite setup connect a data source + configure the model, and start the local stack infinite ask a question infinite "how many page views in the last 7 days" git clone https://github.com/Infinite-Labs-AI/infinite-os.git infinite cd infinite ./infinite setup ./infinite setup installs/builds the workspace if needed, configures model auth, and starts the bundled Docker stack. | Command | What it does | |---|---| infinite | Start an interactive session ask questions in natural language | infinite "