PromptShuttle

PromptShuttle is an OpenAI-compatible API and web dashboard for developers and B2B teams that need server-side multi-agent orchestration without embedding agent logic in application code.

Key capabilities include:

• Multi-agent orchestration in one API call: define agent flows in the dashboard, then call a standard chat completions endpoint to spawn sub-agents, run tool calls, and return an aggregated final result.
• Proxy to multiple LLM providers: route each agent to different models across providers (including OpenAI, Anthropic, Google, DeepSeek) and switch models without code changes beyond base URL configuration.
• Observability and cost controls: visualize the agent execution graph, inspect per-step timing and cost breakdowns, and track usage with budgets per tenant and per flow.
• Team workflows: reusable prompt templates, structured prompting and context engineering, result comparison across models, plus collaboration features for deployed applications.

Pricing: Freemium (free tier with limits, paid plans for advanced usage).

Notable for acting as an API, not an SDK framework, keeping orchestration logic in the proxy so teams can update agents and routing via configuration instead of redeploying application code. It also supports MCP so MCP-compatible clients can manage flows and prompts via a hosted MCP server endpoint.