Help Center

Find answers to common questions and get the most out of your quiva.ai experience

Frequently Asked Questions

What LLM Providers and Models do you support?

quiva.ai supports all major LLM providers through our flexible integration system:

  • OpenAI: GPT-4, GPT-3.5, and all latest models
  • Anthropic: Claude models (including Claude Sonnet, Opus and Haiku)
  • Google: Gemini flash and pro models
  • And more...: more models coming soon

For enterprise plan customers, we also support private model deployments and custom model hosting.

Where do I get an API key for my Agent?

API keys are obtained directly from your chosen LLM provider:

Once you have your API key, you can store it as a Secret inside quiva.ai and reference it inside your agent configuration in the Settings panel. We encrypt and securely store all API keys.

What triggers can I use for ai agents?

quiva.ai offers multiple trigger types to activate your agents:

  • Manual Triggers: Start agents on-demand through the chat trigger interface
  • HTTP Webhooks: Trigger via API calls from external systems
  • Schedule Triggers: Run agents at specific times or intervals
  • Process related triggers: Set up workflow nodes to trigger when emails are received, or files uploaded etc
  • Stream events: React to messages written to event streams

You can combine multiple triggers for sophisticated automation workflows.

How can I orchestrate multiple agents in one workflow?

Our visual Flow Builder makes multi-agent orchestration simple:

  • Sequential Processing: Chain agents to process data step-by-step
  • Parallel Execution: Run multiple agents simultaneously for faster processing
  • Conditional Logic: Route data based on agent outputs and business rules
  • Human-in-the-Loop: Add approval steps where needed
  • Error Handling: Define fallback paths for robust workflows
  • Data Transformation: Format and transform data between agents
  • Composite Workflows: Combine workflows within a single composite flow

Use our drag-and-drop interface to visually design complex multi-agent workflows without coding.

How can I publish something to the Marketplace?

Publishing to the quiva.ai Marketplace is straightforward:

  • Create Your Solution: Build and test your agent, smart MCP interface or whoe agentic workflow
  • Add Documentation: Provide clear instructions and uses cases to help others get the best out of it
  • Set Pricing: Choose between free community or paid premium offerings
  • Submit for Review: Our team reviews for quality and security
  • Go Live: Published items are instantly available to all users

Contributors can earn revenue from paid items. View our publisher guide for detailed steps.

How do I connect my agents to existing tools and integrations?

quiva.ai provides several ways to integrate with your existing systems:

  • Pre-built MCP Servers: Search our Marketplace for pre-made MCP interfaces to popular tools
  • OpenAPI Import: Automatically generate smart MCP server interfaces from API specifications
  • Postman Collections: Convert existing Postman collections to MCP servers
  • Custom HTTP Requests: Add direct API integrations in to to your workflows with a HTTP function call
  • File System Access: Read and write data to the stream-based storage inside your quiva.ai account

Our Smart MCP converter makes it easy to make any API enabled system AI-agent compatible.

Missing Something?

Let us know what other help topics you'd like to see covered