Overview
Monid is a routing layer built for autonomous agents, not humans clicking dashboards. You hand your agent a single skill file (or point it at a remote MCP server, or install the CLI), and it can then call monid.discover() to rank matching tools by fit and price, inspect an endpoint’s schema, and run it — all billed against one prepaid balance. There are no monthly seats; you pay a tiny fee per call (example runs show roughly $0.0013 per lightweight call, plus a 10% platform fee on the underlying provider cost). The catalog leans heavily toward what agents actually need: web scrapers (Apify, Browserbase), search (Exa, Octen), people and company data (People Data Labs, Apollo), AI media generation (ElevenLabs, MiniMax, Seedance), and on-chain or weather feeds. Failed requests are not charged, and workspaces carry budgets with per-run caps so a runaway loop gets BLOCKED rather than draining your wallet. For a solo builder wiring a research bot, an outbound sales agent, or a competitor-monitoring cron, Monid collapses a week of API integration into one line of setup.
Key Features
- Runtime tool discovery — monid.discover() returns ranked candidates with live prices and input schemas, so the agent picks the cheapest capable endpoint.
- Unified pay-per-call billing — one shared balance covers 1,800+ tools across 55+ providers; no separate keys or subscriptions.
- Three integration paths — a one-line skill for chat agents, a remote MCP server for any MCP client, and a CLI/HTTP API for code.
- Spend guardrails — workspace budgets and per-run caps block overspending, and failed provider calls are not billed.
- Provider marketplace — tool owners can list their own APIs and share revenue from metered agent traffic.
Pricing
| Plan | Price | For |
|---|---|---|
| Free credit | $1 on signup | Try the registry with no commitment |
| Pay per call | ~$0.0013–$0.10+/call | Variable agent workloads billed from one balance |
| Platform fee | +10% on provider cost | Applied on top of each underlying call |
| Provider tiers | Flat fee or per-result | Some endpoints bill per result with an optional base fee |
Pricing is consumption-based with no subscriptions. New accounts start with $1 in free credit, and example documented runs range from a few hundredths of a cent for a social feed read to a couple of dollars for a scripted, voiced video. Actual cost varies by provider and billing model.
Comparison
Compared to OpenRouter, which routes model calls to the cheapest capable LLM, Monid applies the same idea to tools — it routes agent actions to the cheapest capable API. Compared to Zapier, which hard-wires fixed integrations through a no-code UI, Monid lets the agent assemble its own toolchain at runtime instead of you pre-building every zap. And versus a full Replit agent that builds and deploys apps, Monid is narrower but far lighter: it is plumbing for calling external APIs, not a place to host code.