How to Use MoltSets: A Step-by-Step Setup Guide

To use MoltSets, you sign up for a free account, pick one access method (MCP for AI tools like Claude, an API key for the REST API or CLI, or the browser extension), run a first lookup, and then check the status and risk_score on what comes back. There is no list-building dashboard, so every step below happens in an AI client, a terminal, Clay or LinkedIn. If you are still deciding whether the tool fits you, start with our MoltSets review.

Some tool links here are affiliate links (disclosure). Nerdy Joe is a MoltSets affiliate. We have not tested MoltSets hands-on; every step below comes from MoltSets' own documentation pages, cited in each section and in the sources at the end.

Before you start

MoltSets calls itself a B2B search and enrichment service with "no UI": emails, carrier-verified mobiles and company data go straight to your sequencer, CRM, AI tool or spreadsheet. Its FAQ says there is no dialer and no sequencer either.

What you need, per MoltSets' docs:

• A MoltSets account (free, no card).
• For MCP: a client that supports remote HTTP MCP with OAuth and PKCE, such as Claude.ai, Claude Desktop, Claude Code, ChatGPT, Cursor, Windsurf or Grok.
• For the CLI: Node.js 18 or later.
• For the extension: Chrome 116 or later, or Firefox.
• A test list you can judge, such as 10 LinkedIn URLs of people whose emails you already know.

Pick your access method

All methods share one account and one set of Fair Use limits. MoltSets says limits are per account, shared across every API key and connector.

Step 1: Create your account

MoltSets' how-it-works page lists three steps:

1. Sign up at app.moltsets.com/signup. No card is needed, and you get free records to test.
2. Stay on Free or choose a paid plan. MoltSets says the only difference between plans is the rate limit.
3. Connect to Claude, Clay, Codex, the CLI or another tool, then search for prospects or enrich an existing list.

The free tier gives about 1,000 lifetime lookups (MoltSets describes it both as 1,000 records and 1,000 API calls) and 100 lifetime search records. That is plenty for the tests in this guide. You can create your free MoltSets account and come back to step 2.

One rule worth knowing: MoltSets says "an agent can't create a MoltSets account." A person signs up first, and the agent connects afterward.

Step 2: Get the right credential

This is where most setup mistakes happen, because MoltSets uses two kinds of credentials.

• API key. Create it under API keys in the MoltSets app. Keys start with ms_, and MoltSets suggests creating several (for example one per environment or agent). The REST API, the CLI and the extension use API keys.
• Connector credentials. Open Connectors in the MoltSets app to get a Client ID and Client Secret. MCP clients use these with OAuth.

The CLI page is explicit: MCP Client ID and Client Secret do not work with the CLI. If a CLI command fails with unauthorized, check that you used an ms_ key.

Step 3: Your first lookup in Claude (MCP)

MoltSets runs a remote HTTP MCP server at https://mcp.moltsets.com/mcp with OAuth. Its MCP page gives these steps.

Claude Code

Run this command, replacing YOUR_CLIENT_ID with the Client ID from Connectors:

claude mcp add --transport http moltsets \
  https://mcp.moltsets.com/mcp \
  --scope user \
  --client-id YOUR_CLIENT_ID \
  --client-secret

MoltSets says the command prompts for your secret, then opens the MoltSets consent screen in your browser. Approve access there, and the MoltSets tools appear in your agent.

Claude.ai, ChatGPT and Cursor

• Claude.ai: add MoltSets as a custom connector with the server URL and your OAuth credentials. On Team or Enterprise, an Owner adds it once for the organization.
• ChatGPT: add MoltSets as a custom connector (Plus, Team and Enterprise, per MoltSets).
• Cursor: add the server from Cursor Settings or in .cursor/mcp.json.

MoltSets links a separate guide for each client, so follow the one for yours for the exact fields.

Test prompt

MoltSets' MCP page suggests this first prompt:

Enrich this LinkedIn profile and give me their business email and phone number: linkedin.com/in/example

Swap in a real profile you can check. A mobile number needs phone tokens on a paid plan, so on Free expect the email only.

The setup prompt shortcut

MoltSets also documents a setup prompt approach: paste a prompt that tells your agent to read https://moltsets.com/llms.txt, follow the Getting Started section, and confirm the connection with get_account. Because get_account is a free account call, this test costs nothing.

Step 4: Your first lookup with the REST API

If you are wiring MoltSets into code, the API page gives this pattern. Every tool is a POST to https://api.moltsets.com/api/v1/tools/ followed by the tool name, with a JSON body.

curl -X POST https://api.moltsets.com/api/v1/tools/linkedin_to_best_email \
  -H "Authorization: Bearer ms_XXXXXXXXXXX" \
  -H "Content-Type: application/json" \
  -H "User-Agent: my-app/1.0" \
  -d '{"linkedin_url": "https://linkedin.com/in/example"}'

Two details trip people up. MoltSets requires a non-empty User-Agent header (connections without one get 403 since September 16, 2026), and a miss returns HTTP 200 with status "not_found," not a 404.

For batches, MoltSets accepts linkedin_urls or emails arrays of up to 100 items, and each item counts as one request. Our MoltSets API guide covers the full tool list, error codes and 429 handling.

Step 5: Enrich a CSV with the CLI

The CLI is the most practical way to enrich a list you already have. These steps follow MoltSets' CLI page.

Install and sign in

npx moltsets --help
npm install -g moltsets

The first command runs without installing; the second installs globally. The CLI reads credentials in this order: the --api-key flag, the MOLTSETS_API_KEY environment variable, then a key saved by moltsets login. MoltSets recommends the environment variable for scripts, CI and agents.

Check that it works with a free call:

moltsets whoami

MoltSets says whoami shows your plan, token balance and Fair Use headroom.

Dry run, sample, full run

Your CSV needs a column of LinkedIn URLs (named linkedin in this example). Run the three steps in order:

moltsets enrich leads.csv --tool linkedin_to_best_email --column linkedin --dry-run
moltsets enrich leads.csv --tool linkedin_to_best_email --column linkedin --limit 10 --out sample.csv
moltsets enrich leads.csv --tool linkedin_to_best_email --column linkedin --out enriched.csv

1. Dry run counts rows and checks your headroom with no paid calls.
2. Sample processes the first 10 rows only. Open sample.csv and compare the emails with what you know.
3. Full run processes the whole file.

What the CLI does for you

• Works in batches of 100 and saves progress after every batch. Rerunning the same command resumes where it stopped; --restart starts over.
• Checks Fair Use headroom before each batch. If the 5 hour window cannot cover it, the CLI waits, or stops and saves with --no-wait.
• On a 429 response, waits the time the API specifies and retries up to four times.
• If tokens run out mid-run, keeps returned results and marks the remaining rows "dropped."
• Keeps all your original columns and adds verified emails, risk scores and a status per row.

If your file has names and company domains instead of LinkedIn URLs, MoltSets lists a --map flag for name and domain columns with search_business_email_by_name. Check npx moltsets --help for the exact syntax.

Step 6: Build a new list with search

MoltSets has no filter screen, so list building happens through the search tools: search_people, search_companies, search_linkedin_profile and others. You can ask your agent in plain language over MCP, call the API, or use the CLI.

The CLI page shows the pattern moltsets followed by a search tool name, with --max 500 or --all to page through results and --out leads.csv to save them. It does not show a full search example, so run npx moltsets --help to see the filter flags before your first search.

What the API reference documents for search_people:

• Filters: title, company, company_domain, country, state, city, seniority, industry, department or functional area, employee range, revenue range, NAICS code and LinkedIn industry.
• Exclusions: company domain, country, state, industry, seniority and employee range. Exclusions cannot be the only filter.
• Paging: offset and limit, with a maximum of 25 results per call.

Two cost tips from MoltSets' FAQ. Run search_linkedin_profile with count_only set to true to see how many matches exist without being charged, and remember that search records have their own Fair Use caps (2,500 per week on Solo). Our MoltSets pricing breakdown has every cap.

Step 7: Read status and risk_score

Every result comes back in an envelope with results, status and metadata. Check status first: "ok" means data came back, "not_found" means no match and no record charged.

Then check risk_score on each email. MoltSets' email risk score docs define five grades:

A few rules from the same page:

• When signals conflict, D wins, then A, B, C and F. Treat a missing risk_score as F.
• An A grade is about send safety, not identity. MoltSets says it "doesn't confirm the address belongs to the person you searched."
• MoltSets suggests suppressing C-grade contacts who do not reply after 1 to 2 attempts.
• Addresses that fail validation are dropped before delivery, so a lookup whose only address failed returns not_found.

A simple sending rule built from those grades: send A, consider B, test C carefully, never send D, and treat F like C. That is our summary of MoltSets' recommendations, not a MoltSets setting. MoltSets publishes no accuracy percentage or bounce guarantee, which our detailed MoltSets review discusses in more detail.

Step 8: One-off lookups with the extension

For a handful of contacts, the MoltSets for LinkedIn extension is quicker than the CLI, and it uses the same MoltSets account. MoltSets' extension page gives four steps:

1. Add MoltSets for LinkedIn from the Chrome Web Store (a Firefox version also exists).
2. Open the side panel, then paste your API key or sign up.
3. Open a LinkedIn profile, choose a lookup (business profile, business email, US personal email or mobile) and click Get.
4. Save to a list, download a CSV, or send the contact to a connected tool such as HubSpot, Attio, Salesforce, Clay, Make, n8n, Zapier or a JSON webhook.

MoltSets says the extension is free and lookups draw on your plan. Misses, reopening profiles, lists, CSV downloads and integration sends are free.

Step 9: Check your usage for free

Because the weekly cap is the one that binds, check usage before a big job. MoltSets says get_account, get_billing and get_usage cost nothing and do not count against rate limits.

• Over MCP: ask your agent to run get_usage for "week."
• Over the API: call get_usage with a period of today, week, month or billing_cycle.
• In the CLI: run moltsets whoami, moltsets billing or moltsets usage.

If you hit a cap, the API returns HTTP 429 with a code such as fair_use_limit_exceeded. Wait for the rolling window to free up, or move to a bigger plan.

Step 10: Get the list into your sequencer

MoltSets does not list a native integration with Instantly, Smartlead, PlusVibe or lemlist. The route that works with documented pieces is:

1. Enrich or build your list with the CLI (--out writes a CSV) or download a CSV from the extension.
2. Filter rows by status ("ok" only) and by risk grade.
3. Import the CSV into your sequencer using its own import steps.

That is our workflow from MoltSets' documented parts, not a MoltSets feature. We walk through it in detail in MoltSets to Instantly. Before sending at volume, make sure your domains and inboxes are ready; our guide to cold email infrastructure covers that.

Common errors and fixes

These codes come from MoltSets' developer docs. MoltSets says nothing is charged on an error.

Setup checklist

• [ ] Signed up and confirmed the plan with get_account or moltsets whoami.
• [ ] Used Connector credentials for MCP and an ms_ key for the API, CLI and extension.
• [ ] Ran a dry run and a 10 row sample before the full file.
• [ ] Filtered on status "ok" and on risk grade before exporting.
• [ ] Checked weekly headroom with get_usage before a big job.
• [ ] Planned the CSV import into your sequencer.

When you are ready to try it on your own data, start with the MoltSets free tier. If you would rather buy a finished list than build one, see our guide to verified lead lists for agencies.

Rather not build the list yourself?

Setting up MoltSets is quick. Building, checking and sending to fresh lists every week is the part that eats your calendar.

Nerdy Joe is owned by Agency Growth Partners (AGP), and AGP builds verified lists and runs outbound for agencies. If you would rather hand that off, book a call with David.

Bottom line

Using MoltSets comes down to three choices: MCP if you work in Claude or another AI client, the CLI if you have a CSV to enrich, and the extension for one-off LinkedIn lookups. Start every job with a free usage check and a small sample, and send only what the risk grade supports.

For pricing, data claims and weaknesses, read the complete MoltSets review. If MoltSets looks like the right fit, sign up for MoltSets and run your first test today.

Sources

• MoltSets how it works
• MoltSets MCP server
• MoltSets AI agents
• MoltSets API
• MoltSets CLI
• MoltSets LinkedIn Chrome extension
• MoltSets llms.txt
• MoltSets developer docs index (llms.txt)
• MoltSets developer docs: API reference
• MoltSets developer docs: rate limits
• MoltSets developer docs: email risk scores
• MoltSets developer docs: Clay integration
• MoltSets integrations
• MoltSets FAQ
• MoltSets homepage
• MoltSets changelog