Maestro Docs

Appearance

Provider Comparison ​

Fresh

Side-by-side feature matrix for all supported AI providers.

Feature Matrix ​

FeatureClaude CodeCodexOpenCodeFactory Droid
StatusPrimaryFullFullActive
Image attachmentsNew + resumedNew onlyNew + resumed-
Session resume--resumeexec resume <id>--session-
Read-only mode--permission-mode plan--sandbox read-only--agent plan-
Slash commandsBatch-mode onlyInteractive TUI onlyNot supported-
Cost trackingFull breakdownToken counts onlyPer-step costs-
Model selectionVia Anthropic account-m, --model--model provider/model-
Context operationsMerge, export, transferMerge, export, transferMerge, export, transfer-
Thinking displayStreaming assistant msgsReasoning tokens (o3/o4)Streaming text chunks-
Mid-turn inputNot availableNot availableNot available-

Execution Modes ​

ProviderCommand Maestro Uses
Claude Codeclaude --print --verbose --output-format stream-json
OpenAI Codexcodex --json (YOLO mode enabled)
OpenCodeStandard execution
Factory Droidfactory-droid -o stream-json

OpenCode Provider Support ​

OpenCode supports 75+ LLM providers including:

ProviderBase URL
Ollamahttp://localhost:11434/v1
LM Studiohttp://127.0.0.1:1234/v1
llama.cpphttp://127.0.0.1:8080/v1

Model ID format: provider_id/model_id (e.g., ollama/qwen3:8b-16k)

Planned Providers ​

ProviderStatus
Gemini CLIPlanned
Qwen3 CoderPlanned

See Also ​

Built with VitePress + RAG Chat

Maestro Documentation