Lawmatics MCP vs Zapier: Two Ways to Connect Claude to Lawmatics

The two tools answer different questions. Zapier answers "when something happens in Lawmatics, how do I push it somewhere else?" An MCP server answers "how does my AI assistant read and work with my intake data in a conversation?" Firms searching for "Lawmatics Claude integration" are usually asking the second question, and Zapier was built for the first.

Lawmatics itself sits in the legal CRM and intake category alongside Clio Grow, Lead Docket, Captorra, and Lawcus. Everything below is specific to Lawmatics and its API, because that is where the details live.

What Zapier's Lawmatics integration actually does

Checked against Zapier's own integration listing as of August 2026, the Lawmatics app offers:

• Triggers: Matter Converted, Matter Added, Matter Updated, Appointment Created, Contact Created, Form Entry Created, Custom Form Added, Note Created.
• Actions: create or update a Contact, Company, or Matter, Create Appointment, Submit Form Entry.
• Searches: find a Contact, Company, or Matter by email, name, or phone.

For relay automation, that surface is genuinely useful. A new website lead becomes a Lawmatics contact. A converted matter fires a Slack message or a QuickBooks entry. A form submission lands a row in a Google Sheet. Zapier handles those handoffs reliably and has for years.

Now look at what the surface skips, because for a Lawmatics account the gaps are unusually painful:

• No custom fields in the create and update actions. Lawmatics intake data lives largely in custom fields, so a Zap can create a contact but cannot carry the answers your intake form actually collected.
• No invoices. Not even read access.
• No pipelines, stages, or sources reads. You cannot ask Zapier where a matter sits in the pipeline or which marketing source produced it, which is the core of what Lawmatics measures.
• No tasks.
• No bulk operations. Zaps process one record per trigger.

None of this is a criticism of Zapier's engineering. It reflects what a trigger-action platform is for. It does mean that "Lawmatics is on Zapier" tells you much less than it sounds like it tells you.

What the MCP server does differently

MCP (Model Context Protocol) is the open standard AI assistants use to work with external systems as tools during a conversation. Our open-source Lawmatics MCP server (@oktopeak/lawmatics-mcp, released August 2026, MIT license) gives Claude 37 tools: matters (which the Lawmatics API calls "prospects"), contacts, companies, pipelines and stages, practice areas, marketing sources, sub-statuses, users, tasks, events, notes, timeline activities, tags, custom fields, intake form submissions, and read-only invoices. It is the first Lawmatics server listed in the official MCP Registry, published as io.github.oktopeak/lawmatics-mcp, and the code is on GitHub.

In practice that means you type questions in plain language: how many open matters sit in each pipeline stage, which marketing source produced the matters that actually converted this quarter, what did this client answer on the intake form, which tasks are overdue and whose are they. Claude picks the tools, pages through results, and assembles the answer. We walk through the reporting patterns in detail in Lawmatics pipeline reporting with Claude.

Three design choices matter for a law firm evaluating it:

• A real read-only mode. Set LAWMATICS_READ_ONLY=1 and every write tool is removed from what Claude can see. This is a hard guarantee at the connector level, and cautious firms should start there.
• A local audit log. Every tool call is appended to a JSONL log at ~/.lawmatics-mcp/audit.log on your machine, so you hold your own record of what the AI touched.
• A client-side throttle of 50 requests per minute by default. Lawmatics documents its rate limit as 50 per minute in one place and 150 per minute in another; the connector defaults to the safe number.

One thing we will say plainly because nobody else will: the connector was built from Lawmatics' official API documentation and has not yet been verified against a live firm account. Lawmatics gates API access behind a support request, and the docs contradict themselves in places (the rate limit above is one example). Everything shipped matches the documented behavior; if your account behaves differently, we want the bug report.

Where Zapier genuinely wins

Three advantages are real and we concede them without hedging.

Zapier reacts to events. The MCP connector does not. The connector is pull-only: Claude asks, Lawmatics answers. It cannot wake up because a form was submitted or a matter converted. If you need something to happen automatically the moment an event fires, Zapier does that today. (Lawmatics also supports 13 webhook event types, including matter.created, form.submitted, and invoice.paid, but those are configured in the Lawmatics dashboard and you need somewhere to point them.)

No code, no process. Zapier is hosted. There is no Node.js requirement, no config file, no laptop process. The MCP install takes about 15 minutes and a JSON edit, which we documented step by step in the Lawmatics MCP setup guide, but it is still an install.

7,000+ other apps. If the job is moving Lawmatics data into Mailchimp, QuickBooks, Slack, and a spreadsheet from one dashboard, that is Zapier's home turf and no single-system connector competes there.

The comparison in one table

A note on Zapier MCP (zapier.com/mcp/lawmatics), since it sounds like it closes the gap: it exposes the same Zapier action catalog to AI assistants, roughly 7 write and 5 search actions for Lawmatics. It inherits every limitation of the classic integration, so no custom fields, no invoices, no pipeline reads, no tasks, and each tool call consumes tasks from your plan quota. It is the same thin surface with a conversational doorway.

The questions firms actually ask us

Can Claude see my intake form answers through Zapier?

Only at the moment of submission, and only forward. Zapier's Form Entry Created trigger fires when a form comes in and can push those answers into another app right then. What it cannot do is look one up later: there is no search or read action for form entries or custom fields. The MCP server reads intake form submissions and custom-field values on demand, on any record, months after the fact. The underlying API supports ?fields=custom_fields and even ?fields=all, and the connector uses both.

Which one handles "how many matters are in each stage, by practice area?"

Only the MCP server, and even there it is honest work. The Lawmatics API returns 25 records per page with no way to change the page size, and accepts exactly one filter per request. So a pipeline-wide report means Claude walks the pages, applies the second condition itself, and totals up. The connector's 50 requests per minute throttle keeps that sweep inside the rate limit. Zapier has no path to this at all: no pipeline reads, no bulk reads, no aggregation.

Can either one move a matter to a new pipeline stage?

Neither, because the Lawmatics API itself has no endpoint for it. Stage moves happen in the Lawmatics UI or through Lawmatics' own automations. The closest API write is updating a matter's sub-status. Any vendor telling you their integration "moves matters through your pipeline" via this API deserves a follow-up question.

What about Lawmatics' own webhooks and automations?

Worth a look before you pay anyone for triggers, because some of Zapier's job is already built into Lawmatics. The platform supports 13 webhook event types, including matter.created, form.submitted, and invoice.paid, configured in the Lawmatics dashboard, and its native automation engine handles a lot of the drip-and-follow-up work firms first reach to Zapier for. One more API detail that surprises people: the form submission endpoint is unauthenticated and fires your Lawmatics automations on submit, which is how embedded intake forms work. If your "integration need" is really "run my intake sequence when a form comes in", Lawmatics may already do it without either tool in this comparison.

What does a month of real use cost on each?

Zapier meters tasks. Every record a Zap touches consumes at least one task, so a firm syncing 60 leads a month with three steps per Zap burns through the free tier's 100 tasks in the first week and lands on a paid plan from around $19.99 per month for 750 tasks, scaling up from there with volume. The MCP connector is $0 at any volume. There is no task meter and no per-seat fee; the only subscription in the loop is the Claude plan you already pay for. If you are doing the wider spend math on the platform itself, we put the numbers in what Lawmatics actually costs.

The client-data question

Lawmatics is a CRM, so most of what flows through it is prospect and intake data rather than privileged matter files, but "less sensitive than the case file" is a low bar and intake answers can be sensitive on their own. The structural difference between the two routes is the same one we walk through for practice management systems: with Zapier, your data and your Lawmatics credentials transit Zapier's cloud, which adds one more company to your confidentiality analysis whether or not it changes your conclusion. With the local MCP connector, data moves between your machine, the Lawmatics API, and your AI provider under your firm's own agreement, and the access token never leaves the config file on your machine.

On the AI provider side the honest framing is the same for both: with any hosted model, whatever the tool reads goes to the provider for inference. The defensible setup for sensitive content is a plan with training disabled and zero data retention, and that choice matters more than which integration route you picked.

Can you use both? Yes, and many firms should

This is the part both vendors' marketing skips: nothing about the two tools conflicts. They authenticate to the same Lawmatics API independently and never interact. A sensible split at a firm running Lawmatics today:

• Zapier keeps the event-driven relays it already runs: new lead into Lawmatics, converted matter into the billing system, appointment into a shared calendar.
• The MCP server takes the questions: pipeline counts, source ROI, intake answers, overdue tasks, invoice status, the Monday-morning "where does everything stand" sweep.

The one shared resource is the Lawmatics rate limit, and at 50 requests per minute on the connector side plus low-volume Zaps, a small firm will never feel it.

The verdict

Use Zapier when the job is an event-driven push from Lawmatics into other apps with no AI in the loop. It fires on events, needs no code, and reaches 7,000+ apps. Nothing else does that job better, including our connector.

Use the MCP server when the job is conversational access to your own data: reporting across pipelines and sources, custom fields, intake form answers, tasks, invoices, and bulk reads. Zapier's Lawmatics surface cannot reach most of that data at any price, and the connector reaches it for $0.

Where Oktopeak fits

We built and maintain the connector, it is free and open source, and it stays free whether or not you ever talk to us. Weave's lawmatics-mcp.com directory lists MCP servers in this space if you want to survey the field yourself. If you would rather not do the install alone, we walk the first person at your firm through it on a screen share, at no cost, as part of our guided MCP setup. Most firms are asking their own pipeline questions by the end of that call. Start with the setup guide if you want to see every step before deciding.

Sources: Zapier's Lawmatics integration listing and Zapier MCP page (zapier.com/mcp/lawmatics), Zapier published pricing, the Lawmatics API documentation at api.lawmatics.com/v1, and the @oktopeak/lawmatics-mcp source on GitHub. Vendor surfaces checked August 2026; both change, so verify against the live listings before you commit.