Monid
Connect your agent to every tool it needs
About Monid
Monid is a metered registry that gives an agent access to third-party APIs at runtime instead of requiring you to wire integrations in advance. The catalogue held 1,947 tools across 67 providers when it was checked, spanning search, SEO, lead generation, scraping, media generation, weather, finance and on-chain data. The shape of the problem it addresses is familiar to anyone who has built an agent: every capability you want means another vendor account, another API key, another subscription and another integration written before the agent can use it, and most of those subscriptions go underused. Monid holds the provider keys and brokers the calls, so the agent sees one balance, one rate-limit envelope and one invoice.
The commercial model is the most interesting thing about it and it is stated plainly rather than buried. Billing is per call with no subscriptions and no minimums, each endpoint publishes its own rate per call or per result, and Monid adds 10% on top of the underlying provider price. New accounts start with a dollar of free credit, and their own FAQ says a request that fails to deliver results is not charged. Set against that, two things deserve weighing before you build on it. There is no pricing page - the model lives in a homepage FAQ answer and the individual rates live inside the catalogue, so estimating a workload means going and looking. And a large share of the catalogue is scraping endpoints for social and commerce platforms, which is a terms-of-service question you own, not the registry. The company is a Delaware C-corp in San Francisco and the site labels itself v0.1.0.
There are three ways in and they share the same registry and the same balance. The lightest is a skill: one line pasted into an agent's chat pointing it at the published skill file, after which the agent handles setup itself. The second is a remote MCP server added to a client that speaks MCP, with the documentation naming Claude Code, Codex, Cursor and OpenCode among the agents that work. The third is a CLI installed and run from a shell, with an HTTP API underneath for calling directly from code. The working loop is discover, inspect, run and pay. A discover call returns ranked candidate endpoints for a described task, ordered by fit and by price, so cost is visible at selection time rather than after the fact. The agent then inspects an endpoint, runs it, and the charge is debited from the single balance. Because Monid holds the upstream credentials there is no per-provider signup step in the middle, and it pools capacity across providers so a throttled provider can be retried against the next-best one.
- •A Registry of Nearly 2,000 Tools - Search, SEO, lead generation, scraping, media generation, finance, weather and on-chain data across 67 providers, all reachable through one integration
- •Discovery Ranked by Fit and Price - A discover call returns candidate endpoints ordered on both, so the agent can choose on cost before it commits
- •One Balance, No Subscriptions - Pay per call with no minimums, with a stated 10% margin over the underlying provider price and no per-provider accounts to maintain
- •Three Ways to Connect - An agent skill, a remote MCP server, or a CLI with an HTTP API underneath, all against the same registry and balance
- •No Provider Keys to Manage - Monid holds the upstream credentials and brokers the calls, presenting one rate-limit envelope and one invoice
- •Pooled Capacity and Failover - Requests retry against the next-best provider when one is throttled, and their FAQ states failed requests are not billed
Developers building agents that need to reach the outside world and who do not want to pre-commit to a stack of subscriptions to find out which capabilities matter. The pay-per-call model fits exploratory and bursty workloads better than a set of monthly plans, and the discovery step is genuinely useful when you do not know which vendor to use for a task. Anyone already running Claude Code, Codex, Cursor or OpenCode can attach it without writing an integration. Think harder before adopting it if you need to forecast costs precisely, since there is no pricing page and per-endpoint rates only appear inside the catalogue; if brokered credentials are a problem for your security review, since the upstream keys are held by the vendor; or if your use would lean on the scraping endpoints, where the terms of service exposure is yours to carry. The v0.1.0 label is a fair description of the maturity on offer.
Pricing
Usage based, no flat monthly plan
Pay-per-call with no subscriptions and no minimums. Each endpoint publishes its own rate per call or per result and Monid adds 10% on top of the underlying provider price. New accounts start with $1 in free credit and their FAQ states failed requests are not charged. There is no pricing page - this is from the FAQ on their homepage.
From the vendor pricing page, 2026-10-16
