FAQ for Developers

What is CableKnit?

CableKnit is an AI automation marketplace and runtime. You build plugins that define workflows, skills, and data integrations — the platform handles execution, hosting, billing, and delivery. Your code runs on our runtime infrastructure, so businesses install your plugin and it just works — no servers to manage, no infrastructure to maintain. When companies subscribe, you earn monthly recurring revenue with payouts processed automatically.

What kind of plugins should I build?

CableKnit is built for small and medium-sized businesses in niche industries — the companies that need automation but don't have engineering teams to build it themselves. The best plugins solve real operational problems for a specific field: invoice routing for accounting firms, patient intake for dental offices, permit tracking for contractors, inventory alerts for independent retailers.

The ideal CableKnit developer is someone who knows an industry well. If you've worked in insurance, logistics, property management, veterinary practice, or any specialized field, you already understand the workflows that waste people's time. That domain knowledge is your edge — you know which problems are worth $500-$2,000/month to solve because you've lived them.

You don't need to be a full-stack engineer. CableKnit plugins are JSON configuration — workflow definitions, skills, and data sources — not traditional application code. If you can describe a business process step by step, you can build a plugin. The CLI and MCP tools handle the technical details.

How do I get started?

Apply for a developer account, then install the CableKnit CLI:

The CLI walks you through scaffolding a plugin, validating it, and pushing it to the marketplace. Run cableknit --demo to explore the full workflow with mock data before writing a line of code.

How do I publish a plugin?

Use the CLI to scaffold, build, and push your plugin bundle. A bundle includes a plugin.json manifest, automation workflows, skills, artifact blueprints, data source tools, documentation, and images. Run cableknit validate to check for errors, then cableknit push to publish. Your plugin goes through a quick review before appearing on the marketplace.

How do I make money?

You set a monthly price for your plugin (minimum ~$500/month depending on AI usage). Companies subscribe and pay monthly. CableKnit handles all billing and processes payouts via Stripe Connect.

Connect your Stripe account from the developer portal, then track your published apps, subscriber count, and year-to-date earnings from the earnings page or the CLI.

Why build on CableKnit now?

CableKnit is a new marketplace and runtime, and both are early. That means less competition, more visibility, and a real chance to establish yourself before the space gets crowded.

Early developers have advantages that disappear at scale:

• First-mover positioning in underserved industries — if you're the only invoice automation for logistics, every logistics company on the platform sees your plugin
• Direct access to the team — we review every submission personally and actively help early developers ship
• Revenue compounds — plugins earn recurring monthly revenue, so launching early means more months of compounding subscribers
• Your plugins shape the platform — early developers influence what connectors, platform tools, and features we build next

Every marketplace has a window where early builders disproportionately benefit. Shopify app developers who launched early, Stripe partners who built integrations before the ecosystem matured, Slack apps that got in before the directory was saturated — the pattern repeats. The best time to build is before everyone else shows up.

What can a plugin do?

Plugins combine AI assessment, human-in-the-loop decisions, email and notification actions, external service integrations via connectors, and data processing. For example, a plugin could automatically triage incoming invoices, route high-value ones to a human for approval, and notify accounting when done — all without the company writing any code.

Automations can be triggered by schedules, inbound emails, connector events, or webhooks — so your plugin can react to the real-world events that matter to the business.

What are platform tools?

Platform tools are shared capabilities your plugin can use at runtime via LLM function calling — things like fetching briefings, requesting human approval, transforming files, or updating spreadsheets. Declare the tools you need in plugin.json and the platform provides them. Browse available tools with cableknit tools list.

What are connectors?

Connectors are pre-built integrations with external services like Slack, GitHub, Salesforce, Google Sheets, and more. Your plugin can read and write data through connectors without managing OAuth flows or API credentials — the platform handles authentication for each company that installs your plugin. Browse available connectors or run cableknit connectors list from the CLI.

Need a connector we don't have yet? Request one from the developer portal — tell us the service, link to API docs, and describe the capabilities you need. We use these requests to prioritize what we build next.

Can I use AI to help me build plugins?

Yes — the CableKnit CLI includes a built-in MCP (Model Context Protocol) server that works with Claude Code, Cursor, Windsurf, and other AI-powered editors. Run cableknit mcp install and your editor's AI assistant gets access to tools for schema lookups, code generation, validation, error explanation, workflow design, and more.

Ask your AI assistant to generate a plugin.json, scaffold an automation, explain a validation error, or review your bundle before submitting — it has full knowledge of the CableKnit platform baked in.

Can I roll back a bad push?

Yes. Every cableknit push stores a copy of the bundle on the server. If something goes wrong, roll back to any of your last 10 versions:

The rollback re-validates the bundle against current platform rules and re-ingests it. Your plugin's approval status is preserved — a live plugin stays live. Run cableknit versions my-plugin to see your version history.

How do I monitor my plugin after it's installed?

The CLI gives you full observability into your plugin's activity across all installs — logs, metrics, and error tracking — without exposing any company-specific data.

• cableknit logs <slug> — view a chronological stream of events (sandbox executions, automation runs, bundle pushes, errors)
• cableknit metrics <slug> — see tool call counts, error rates, automation stats, average sandbox duration, and active installs
• cableknit errors <slug> — quickly surface just the errors with expanded metadata

All commands support --json for scripted output and filtering flags like --type, --severity, and --since. Logs are retained for 90 days.

What is the developer portal?

The developer portal is your home base on the web. From there you can view your dashboard, manage plugins and submissions, track earnings, browse platform tools, request new connectors, and access your sandbox — all without touching the CLI.

How does the sandbox work?

Every developer account includes a sandbox environment that simulates a real company installation. You can trigger automations, view runs, test decision queues, and inspect artifacts — all without affecting production data. The sandbox resets cleanly so you can iterate quickly.

Where are the docs?

The CLI includes built-in documentation — run cableknit and select "Read Me" or "View Sample Plugin" to read guides and a fully annotated example. For the full reference, visit the developer documentation.

What are the app review guidelines?

Every plugin goes through a review before it appears on the marketplace. We check for working automations, documentation, accurate descriptions, and fair pricing. Read the full App Guidelines for details on what we look for and what gets rejected.