MiMo Code
Product information, use cases, and access for MiMo Code.
Pricing information
No verified public pricing is available yet.
DevPrice organizes public information and does not sell MiMo Code subscriptions. Prices and availability are determined by MiMo Code.
What is MiMo Code?
MiMo Code is an open-source AI coding assistant from Xiaomi’s large-model team, designed for developers who want help understanding, changing, and delivering work in software projects. Built on OpenCode, it is primarily presented as a terminal-oriented coding experience where natural-language requests can be turned into development tasks.
The product combines conversational assistance, code editing, and project memory for ongoing work on a codebase. Its Compose mode brings design, planning, coding, testing, and review into one workflow, while persistent memory and the Dream command help retain and organize project context over time; voice input and connections to third-party models provide additional ways to interact with the assistant.
Key features of MiMo Code
Persistent Project Memory
The system preserves working context through project memory, conversation checkpoints, and task progress, with an independent subagent helping record that state. This is suited to coding work that continues across long conversations and multiple stages.
Compose Workflow
Compose mode starts from a development idea and carries it through design, planning, coding, testing, and review. It fits projects where users want several stages of implementation coordinated within one workflow.
Dream Memory Consolidation
The built-in Dream command periodically organizes earlier conversations by consolidating memories, removing duplication, and validating paths. It is intended for long-running development work where project context is accumulated over time.
Voice and Multi-Model Interaction
MiMo Code supports voice input for prompts, instruction changes, and execution commands, while also allowing connections to third-party models beyond MiMo. This gives developers another interaction method and flexibility in choosing a model for different tasks.