Local desktop companion for OpenAI Codex that manages accounts, routing, sessions, MCP entries, skills, relay providers and local configuration.
AiMaMi is an independent open-source desktop companion for OpenAI Codex. It consolidates account and quota management, automatic account switching, session inspection and cleanup, smart routing and relay-provider configuration, MCP and Skills management, custom instructions, diagnostics and local Codex maintenance in one application. It is built with Tauri 2, React and Rust and works directly with Codex data under the local Codex home rather than replacing Codex itself.1
Current platform support is strongest on macOS and Windows. The repository documents universal macOS support and Windows x64 installers, while Linux is described as best-effort for some workflows.1 Smart-router and relay features run through AiMaMi's local proxy, so the application needs to remain active while those routes are being used.1
AiMaMi is Apache-2.0 licensed and available as free open-source software.1 The software itself therefore has no subscription price, but users still bear any costs associated with Codex, third-party model providers, relay services and their own operating environment.
The project publishes packaged macOS and Windows releases in addition to source builds. The latest release visible during this review is version 1.2.6, released on August 2, 2026, with macOS Universal and Windows x64 assets.2 Linux users should expect more manual setup and less complete workflow coverage than the two primary packaged platforms.
The strongest evidence is transparent first-party material: source code, license, architecture, release history and issue tracking are public. The project also explicitly states that it is independent and not affiliated with or endorsed by OpenAI.1 That transparency is a positive trust signal, but it does not substitute for independent reliability testing.
The canonical 50-source panel audit found no current exact-product review or numeric rating suitable for independent quality validation. Missing coverage is treated as absent evidence rather than a negative score. Exact-product GitHub issues do provide useful qualitative compatibility evidence: issue #69 reports protocol-selection problems for Responses-API-capable upstream models, while issue #99 reports a `codex_apps` MCP startup problem under an AiMaMi-managed proxy configuration.34 These are user-reported cases, not universal failure rates, but they illustrate the maintenance risk of a companion that interacts closely with Codex configuration and routing behavior.
AiMaMi is best suited to developers and Codex power users who regularly manage several accounts, relay providers, MCP servers, Skills and local sessions and who are comfortable maintaining an open-source desktop utility.1 It is especially useful when manual edits across Codex TOML, JSON, SQLite and authentication files have become repetitive or error-prone.
It is less attractive for users who only use a single Codex account with default routing, or for teams that require a vendor-managed support contract and independently validated stability. Users relying on relay models or private Codex behavior should also be prepared to keep both Codex and AiMaMi current as compatibility changes land.
breadth, local control and cost: AiMaMi brings account switching, quota handling, sessions, MCP, Skills, routing, relays and diagnostics into one open-source desktop surface.1 Packaged macOS and Windows releases lower the installation barrier compared with editing Codex configuration entirely by hand.2
Weaknesses follow from the same deep integration. Some workflows depend on Codex implementation details and protocol behavior, so upstream changes can create breakage that requires a new AiMaMi release. Current open issues around Responses API routing and `codex_apps` MCP startup show concrete compatibility edges.34 Linux support is also less complete, and the lack of substantive independent product reviews keeps confidence in the rating LOW.
Not separately stated in the source review.
Products matched by shared tasks, audience, workflow, product type and capabilities.