MoltSets gives you two ways to reach the same identity resolution and contact data: a REST API and an MCP server. This page explains how they relate, when to use each, and how billing, limits, and tools work across both.
Choosing between the API and the MCP
What is the difference between the MoltSets API and the MCP?
They are two front doors to the same product. The REST API is for code: backends, scripts, and automation platforms call it over HTTPS. The MCP server is for AI assistants and agents: it exposes the same MoltSets tools so a model like Claude can call them directly in a conversation. Both run on the same account, the same data, and the same credit balance. Every tool available in one is available in the other.
Do I connect Claude using the API or the MCP?
Use the MCP. The MCP server is built for AI clients, so once you add it to Claude, Claude can call MoltSets tools as part of a chat without you writing any request code. The API is the right choice when you are writing your own software that talks to MoltSets directly, rather than working through an assistant.
Which one should I choose for my use case?
Use the MCP when you want an AI assistant to do the work, for example asking Claude to find a contact and enrich it with a business email inside a conversation or an agent workflow. Use the API when you are building your own product, running a backend job, or connecting a no-code platform such as Clay or n8n that speaks HTTP. If you are not sure, the MCP is the faster way to start, because there is nothing to build.
Can I use both at the same time?
Yes. Many accounts run the MCP for ad hoc work inside Claude and the API for production pipelines. Both draw down the same credit balance, so you can track all consumption in one place through get_usage and get_billing.
Connecting
How do I get an API key?
Generate a key from your MoltSets account. Keys follow the format ms_ followed by a string of characters. Treat the key as a secret, since anyone holding it can spend your credits.
How do I authenticate API requests?
Pass your API key as a Bearer token in the Authorization header:
Authorization: Bearer ms_XXXXXXXXXXX
Every endpoint is a POST request with a JSON body, sent to the base URL https://api.moltsets.com/api/v1/tools. For example, a call to resolve a LinkedIn profile to a business email goes to https://api.moltsets.com/api/v1/tools/linkedin_to_business_email.
How do I connect the MCP to Claude?
Add MoltSets as a connector using the MCP server URL https://mcp.moltsets.com/mcp, then authorise it with your MoltSets account. Once connected, the MoltSets tools appear to Claude automatically and it can call them during a conversation. The same connection works in other MCP-capable clients. For more information, read the MoltSets MCP guide.
Does the MCP need an API key too?
The MCP uses the same MoltSets account and the same credit balance as the API. You authorise the connector with your account when you add it, so you do not paste raw request code or manage headers yourself. Usage through the MCP is metered against your balance exactly as API usage is.
Usage
Can I enrich a whole list of LinkedIn URLs in one call?
Yes. The enrichment tools accept a batch parameter for up to 100 values per call. For LinkedIn tools, pass linkedin_urls as an array instead of the single linkedin_url. For email lookups, pass emails as an array. You cannot mix the single and batch parameters in the same request.
Can I check how many results a search will return before spending credits?
Yes. search_linkedin_profile accepts count_only. Set it to true to get the number of matching results without fetching any data and without charging credits. Use it to gauge whether to tighten or loosen your search before you commit to a full lookup.
What are the rate limits?
Limits apply per account and come in two parts: a per-minute request limit and a daily results limit, both set by your plan. Every call counts toward the request limit, including empty results and account calls. Detailed information about rate limits can be found here.
Credits and billing
Am I charged if no data is found?
No. You are only charged when a call returns data. If a lookup finds no match, the credit cost for that call is zero. In an API response you can confirm this in the metadata object, where tokens_charged reflects what the call cost and tokens_remaining shows your balance afterwards.
Are account and usage calls free?
Yes. get_account, get_billing, and get_usage do not consume credits. You can poll them as often as you need to track balance and consumption.
How do I check my balance and usage?
Call get_billing for your plan, current balance, and per-tool costs. Call get_usage with a period of today, week, month, or billing_cycle for a day-by-day breakdown of calls made, credits spent, successful lookups, empty results, and failures, along with period totals.