FAQ
Frequently asked questions about Runtype — from getting started to billing, building, and beyond.
Getting Started
Do I need my own AI API keys?
Nope! Runtype includes platform keys so you can start building right away — no setup, no accounts with other providers. When you're ready for more control, you can bring your own keys (BYOK) on the Growth plan and above. This gives you direct provider billing, higher rate limits, and access to the full range of models.
Tip: The Build tier includes a $20 one-time inference credit, so you can explore different models without worrying about cost.
How long does it take to build a working Flow?
Most people have a simple Flow running in 5–10 minutes. More complex, multi-step workflows typically take 30–60 minutes. The Quickstart guide walks you through a complete example in under 10 minutes.
Tip: Check out the built-in templates when creating a new Flow — they give you a head start with pre-configured steps for common use cases like support Agents, data pipelines, and more.
Can I use Runtype without coding?
Yes — most of the platform uses visual editors and configuration panels. You can build Flows, create Products, deploy Surfaces, and run Evals all without writing code. For advanced use cases like custom data transforms or runtime tools, some JavaScript is involved, but the editor provides helpers and examples to guide you.
Pricing & Billing
What's included on the Build tier?
The Build tier is free and designed to let you explore everything Runtype offers:
50 Flow executions per day at full speed — after that, requests continue in slow mode (a short delay) so your work keeps running
10 development Surfaces to experiment with different deployment channels
$20 one-time inference credit to try different AI models
Unlimited team members — invite anyone
1-day log retention for same-session debugging
It's a great way to learn the platform, build prototypes, and test ideas before going to production.
How does billing work for AI models?
When using platform keys, AI model usage is covered by your inference credits — you'll see estimated costs in your Usage dashboard so you always know where you stand. When using BYOK (Growth plan and above), you pay your AI providers directly at their standard rates. Either way, you're always in control of your spending.
Tip: You can set a spend cap on paid plans to make sure you never get a surprise bill. Find it in Settings → Billing.
Can I cancel anytime?
Yes — cancel anytime from Settings → Billing. You'll keep your current plan's features through the end of your billing period, then your account moves to the Build tier. Your Flows, Records, and configuration are preserved so you can pick back up whenever you're ready.
Flows & Agents
What's the difference between a Flow and an Agent?
Flows are step-by-step pipelines where you define exactly what happens at each stage — prompts, data transforms, API calls, conditionals, and more. They're ideal for predictable, repeatable workflows.
Agents are conversational AI that decide which tools to use and when, based on context. They're ideal for open-ended tasks like answering questions, researching topics, or handling support conversations.
Tip: Not sure which to use? Start with an Agent for conversational use cases (like a support bot) and a Flow for automation use cases (like processing form submissions or running data pipelines).
Can I reuse a Flow across multiple Products?
Yes! A single Flow can serve as a Capability in multiple Products. Any changes you make to the Flow automatically apply everywhere it's used — no need to update each Product individually.
How do I make Flows faster?
A few strategies that help:
Choose faster models — models like GPT-4o-mini or Claude Haiku are significantly faster for straightforward tasks
Keep prompts focused — shorter, clearer prompts reduce processing time
Optimize data transforms — only process the data you need at each step
Use Evals to compare — run the same Flow with different model configurations and see which gives you the best balance of speed, quality, and cost
Products, Surfaces & Capabilities
What's the relationship between Products, Surfaces, and Capabilities?
Think of it this way:
A Product is what you're building (e.g., a customer support assistant)
Capabilities are the things your Product can do (e.g., answer questions, look up orders) — each backed by a Flow or Agent
Surfaces are where your users interact with it (e.g., a chat widget on your site, a Slack bot, an API endpoint)
One Product can have multiple Capabilities and multiple Surfaces, all working together.
Can I have multiple Surfaces on one Product?
Absolutely. One Product can have a chat widget, an API endpoint, a Slack integration, an MCP server, and more — all accessing the same Capabilities. Build once, deploy everywhere.
How do I update a deployed Product?
Edit the underlying Flow or Agent, test it, and your Product automatically uses the updated version. There's no separate deploy step — changes take effect as soon as you save.
Can I use my own domain for the chat widget?
The chat widget embeds directly on your domain via a snippet. API calls are handled by Runtype's infrastructure. For questions about custom API domains, reach out to our sales team.
Records & Data
What are Records?
Records are structured data objects in Runtype. They have a type (like "customer" or "support_ticket"), a name, and a metadata field where your actual data lives as JSON. Flows can read, create, and update Records — making them perfect for storing context, tracking state, or building knowledge bases.
Can I import existing data?
Yes. Upload CSV files to create Records in bulk. During import, you map your columns to metadata fields so everything lands in the right place.
How does search work with Records?
Runtype supports both keyword and semantic search across your Records. Semantic search finds conceptually related content — not just exact keyword matches — which is especially useful for knowledge bases, FAQ lookups, and support workflows. For best results, keep your Record data clean and well-structured.
Evals & Testing
What are Evals?
Evals let you test and compare how your Flows perform across different configurations. You can run the same Flow against a set of Records with different models, prompt variations, or temperature settings — then compare the results side by side.
They're a powerful way to optimize for speed, cost, and quality before going to production.
How many Evals can I run?
Eval limits vary by plan: 3/day on Build, 10/day on Startup, 50/day on Growth, and 200/day on Team. Enterprise plans have unlimited Evals.
Integrations
Which integrations are available?
Runtype connects to Slack, Google Workspace, Linear, GitHub, Weaviate, Firecrawl, Exa, PostHog, and more. You can see the full list and configure connections in Settings → Integrations.
You can also connect custom MCP servers and build runtime tools that call any external API.
Can I request new integrations?
Yes! Email us at [email protected] with your request. Enterprise customers can work with our team to build custom integrations.
Security & Privacy
Is my data secure?
Yes. All data is encrypted in transit (TLS) and at rest. Sensitive fields like API keys use AES-256-GCM encryption. SOC 2 compliance is in progress. Visit runtype.com/security for the latest details.
Do you train AI models on my data?
No — Runtype never trains models on your data. When using BYOK, your data goes directly to providers using your own keys. When using platform keys, we use providers' zero-data-retention policies so nothing is stored or used for training.
Can I delete my data?
Yes. You can delete individual Flows, Records, or your entire account from Settings. Deleted data is permanently removed within 30 days.
Technical
Is there an API?
Yes. Runtype has a full REST API for triggering Flows, managing Records, configuring Products, and more. You can also use the Runtype SDK for a streamlined developer experience.
What's the default execution timeout?
The default timeout is 30 seconds, configurable up to 300 seconds (5 minutes) per execution.
Can I schedule Flows to run automatically?
Yes! Scheduling is available on all plans. Build supports monthly schedules, Startup supports daily, Growth supports hourly, and Team supports per-minute scheduling. Set up schedules from the Surface configuration panel.
Can I self-host Runtype?
Self-hosting isn't available on standard plans. Enterprise and Platform customers can discuss dedicated infrastructure options with our sales team — contact us to learn more.
Support
How do I get help?
Search these help docs — you're in the right place!
Email us at [email protected] for direct assistance
Enterprise customers get a dedicated Slack channel and priority support
Do you offer onboarding or training?
Enterprise customers receive dedicated onboarding and training. Other plans can schedule paid training sessions — contact [email protected] for details.
Can't find your answer? Email [email protected] or search this help center.