ENZO bundles 2,000+ free AI models into one local, open-source desktop app

A developer has released ENZO, an open-source desktop application that pulls more than 2,000 freely available AI models into a single interface for chat, coding and research. The project runs entirely on the user's own hardware and is pitched at people who want wide model access without keeping track of a dozen separate API dashboards. It is the author's first open-source release, published on the Show HN board.

The centerpiece is an agents tab. A user types a task in plain language and the app builds a purpose-scoped agent for that job rather than routing everything through one general chat window. ENZO also links to Gmail, Google Drive and Calendar, so a user can request a summary of the day's most important mail or have recruiter messages answered with drafts assembled from data stored locally. A coding mode adds a live preview pane for running, inspecting and editing code as it changes.

Security is where the pitch gets concrete. API keys sit in a dedicated vault encrypted with AES-256, which the developer says keeps credentials away from intermediaries, and because the program executes on-device, users can inspect what the backend is doing. An optional password lock seals the platform behind a pass screen and generates a backup key. The author is upfront that this is not airtight: a malicious browser extension or other third party can still lift login tokens, so real-world safety depends on how the machine is used.

ENZO reaches past productivity work. A music tab plays tracks, and a background process runs a language model over listening habits to suggest similar music without being asked. That kind of ambient personalization is uncommon in an open-source client, though it raises the same question the Gmail integration does: how much local data the app keeps and how a user audits or clears it.

The release lands in a crowded field of local AI front ends, where Ollama, LM Studio and Open WebUI already compete for the same users. ENZO's differentiator is breadth rather than depth, trading a curated experience for sheer model count and bundled connectors, and the author concedes the codebase carries bugs typical of a first project. Whether aggregation holds up will depend on how quickly those rough edges get filed down.

What to watch is the feedback loop. The developer has asked openly for criticism, and early responses will likely stress-test the vault's claims, the reliability of the Gmail and Calendar links, and how gracefully the app handles an expired key or an offline model. The repo's issue tracker over the next few weeks will say more about ENZO's future than the launch post does.