Q
Windows support?
Not yet, the Mind daemon speaks Unix sockets (Linux/macOS). Windows timeline is undecided; the site states the limit plainly.
Docs · Install Anvaya
Install Anvaya.
From zero to first remembered session.
Direct answer
Anvaya is pre-1.0 and builds from source: Rust 1.75+ plus Ollama on Linux or macOS. Three moves (runtime, model, agent) then anv init scaffolds your first project in under 50ms. Total time: about fifteen minutes plus one model download.
Install Anvaya
Steps
011 · Install Rust + OllamaRust 1.75 or newer via rustup; Ollama from ollama.com. Verify with rustc --version and ollama --version.
022 · Pull a modelollama pull devstral (default). Confirm it answers on localhost:11434. Larger coders help on refactors; devstral covers daily work.
033 · Build AnvayaBuild the source you were given access to with cargo build --release. One static Rust binary (14–19 MB) plus the memory daemon.
044 · Init your repoRun anv init in your project: .anvaya/{config.json, backups/, sessions/, mind/} appears idempotently. First frame in ~12 ms (measured 2026-08-14).
055 · First sessionLaunch anv, pick a small task, approve tool calls inline. End the session, the graph records its first typed nodes. Session two can retrieve them instead of re-reading.
Questions
Asked About Install Anvaya.
Q
How much disk/RAM?
Binary 14–19 MB (build-dependent), idle ~5 MB, graph grows with your history (SQLite + flat index, typically megabytes). The model dominates: 7B needs ~8GB RAM, 30B-class wants 24GB+.
Q
Do I need API keys?
No on the Ollama path, local models need no account or key. Anthropic and OpenAI-compatible providers are BYOK when you escalate a task.
Run Agents That Fit On Your Laptop.
25.6 MB median RSS. 25 agents ran in parallel on a Core 2 Duo with 4 GB RAM. Hundreds on your machine. Zero cloud required on the Ollama path.
Requires Rust/cargo to build from source. Linux and macOS today, Windows not yet supported. Pre-1.0, public beta. Pricing TBD.