MaveCode

Feature deck

MaveCode capabilities

12 feature groups

MaveCode is an AI development extension for VS Code built around one clear product promise: help teams move faster from brief, design, and existing code to production-ready output while keeping control, review, and brand-specific rules inside the workflow.

A full agentic coding assistant inside VS Code Email-specialist personas Proprietary Figma-to-email bridge Managed backend API configuration

01 / Models

Broad API configuration support

Teams can use premium models for complex builds and can also connect free or lower-cost APIs where available. MaveCode is not locked to one model vendor, so users can choose the model based on task complexity, budget, compliance, and reliability.

MaveCodeAmazon BedrockAnthropicBaseten Claude OauthDeepSeekFireworks AIFriendli GCP Vertex AIGoogle GeminiKenariKimi Code LiteLLMLM StudioMiniMaxMistralMoonshot OllamaOpenAIOpenAI - ChatGPT Plus/Pro OpenAI CompatibleOpencode GoOpenRouterPoe Qwen CodeRequestySambaNovaUnbound Vercel AI GatewayVS Code LM APIxAI (Grok) Xiaomi MiMoZ.ai

02 / Workspace

Reads, writes, and runs commands

  • Read project files and understand repository structure.
  • Write and update files.
  • Refactor code across multiple files.
  • Run terminal commands.
  • Execute build, lint, and test workflows.
  • Inspect command output and iterate on failures.

03 / Integrations

MCP connectivity

Installed MCP servers can connect MaveCode to approved external tools, data sources, or internal workflows without hardcoding every integration into the core product.

Documentation lookupDesign systemsTicketing Testing toolsCustom services

04 / Behavior

Persona system

The benefit is quality control. Instead of one generic AI behavior for every task, MaveCode can apply the right constraints for the job.

Standard email personaEnphase email personaNormal code persona Ask persona

06 / Teams

MaveCode backend API configuration

  • Centralized provider configuration.
  • Consistent model access across users.
  • Reduced setup friction for new team members.
  • Server-held provider authorization.
  • Easier governance over available models.

05 / Design bridge

Proprietary Figma plugin for email builds

MaveCode uses the Figma design as the visual source of truth, while the active persona decides how to convert that design into email-safe code.

  1. Read selected Figma frame
  2. Prepare design for email generation
  3. Detect images, vectors, and icon-style elements
  4. Export deterministic 1x and 2x assets
  5. Export reference image and structured metadata
  6. Support exact mcimg layer naming for image blocks

07 / QA

Email on Acid integration

  • Add Email on Acid API key and password in settings.
  • Store credentials securely in VS Code Secret Storage.
  • Validate credentials.
  • Read active HTML email documents, including unsaved changes.
  • Create a rendering test and open the resulting test URL.

08 / Control

Controlled automation and auto-approval

Users can auto-allow selected actions or require approval each time. Configurable behavior can cover reading files, writing files, running commands, using tools, and workspace operations.

09 / Email

Standard persona for production HTML email

Mobile responsivenessTable layoutsOutlook compatibility VMLDark modeApproved CSSCTA handling Final QA

10 / Builds

Code persona for larger builds

The general code persona can support landing pages, full websites, software projects, and broader agentic coding tasks. High-end models are preferred for stronger reasoning, longer context handling, multi-file planning, and reliable debugging.

11 / Rules

Local rules for folder workspaces

Use .mave rule folders to add project-specific or user-level instructions. Generic project rules live in .mave/rules/rules.md, and persona-specific project rules can live in folders like .mave/rules-code/.

12 / Access

Google sign-in and restricted access

  • Controlled onboarding.
  • Reduced unauthorized access risk.
  • Centralized user identity checks.
  • Cleaner team deployment model.
  • Better alignment with governance expectations.