Skip to content

Integrations

Provider choice without losing local control.

ReplicaMD brings multiple AI runtimes into one desktop workflow while keeping repository state on your machine.

Capability index06 modules
Account-backed coding agents
Additional agent runtimes
Routed models
Local models
Git worktrees
Future provider adapters

Account-backed coding agents

Codex and Claude Code use their official local CLIs and keep authentication in provider-owned account directories.

OpenAI Codex
Claude Code
Separate account profiles

Additional agent runtimes

OpenCode and Gemini CLI run through the same workspace, worktree, and terminal model as the primary coding agents.

OpenCode
Gemini CLI
Consistent session surface

Routed models

OpenRouter and provider-specific keys can power model options such as Kimi, MiniMax, and GLM behind explicit local configuration.

OpenRouter
Kimi and MiniMax
GLM via Z.ai or OpenRouter

Local models

Ollama keeps model execution on infrastructure you control while ReplicaMD supplies the surrounding agent workspace.

Configurable Ollama host
Local model selection
No cloud account required

Git worktrees

Repository isolation is built on standard Git worktrees rather than a proprietary remote workspace format.

Local branches
Reviewable diffs
Independent test runs

Future provider adapters

Calendar sync, plugins, and remote APIs remain planned boundaries. They are not presented as active integrations today.

Provider-neutral contracts
Scoped authentication
Explicit release status

Make it durable

Bring your knowledge and agent work into one system.

Request early access