Category: Product

When we started ModulawAI, it did one thing. It helped you find the law. You could search case law and get a straight, well-cited answer, and that was the whole promise. And you trusted us with it, more of you than I dared hope when we began, and I am more grateful for that than I can easily put into words.

Then you kept asking for more. Draft this. Manage this matter. Remember this deadline. Send this to the client. So we kept building. Today Modulaw researches across jurisdictions, drafts, reads your documents, runs workflows, takes calls, and connects to the tools you already use.

Our mission now is bigger than a search box. We want Modulaw to become your firm’s legal operating system, the single place your practice runs on, from the first client call to the final invoice.

What we have been building for you

Over the last while we have spent our nights on the parts of Modulaw you cannot always see. Three of them matter most right now.

• Your credits go dramatically further. We rebuilt how the assistant reuses context, so the heavy part of each turn is now served from cache. A conversation that used to drain your balance quickly can now go several times further on the same credits.
• You only pay for what is genuinely new. When the assistant re-reads context it has already seen, we bill that at a fraction of the cost and pass the saving straight to you. Fairer by design, not by accident.
• Bigger monthly credit pools. Every plan now carries more monthly credits than before, so there is more headroom for real work before you ever think about a top-up.
• One assistant for the whole matter. Research, drafting, document reading, voice, case management, workflows, and a programmable API, all drawing from a single balance with no per-tool surcharges.

Why our pricing is changing

An operating system is a different kind of promise than a search box. It has to be fast, reliable, and there for you every single day, which takes better models, stronger infrastructure, and a team that actually answers when you need one.

The honest truth is that the only way to build that, and to still be standing beside you in ten years, is to price it sustainably. So we are updating our plans. I will not dress it up. For new subscriptions, some plans now cost more than they did a year ago. We are not doing it to extract more from you. We are doing it to fund the mission you have been asking us to pursue.

And the value moved first. The pricing follows a product that is already faster, more capable, and stretches your credits further than the day you signed up.

For the people already with us

• Your existing workspace is protected. The team seats you have today stay yours, and nobody gets pushed over a limit.
• More credits, not fewer. These changes only ever add to your monthly balance.
• 20% off when you pay annually, on every plan and in your own currency.
• No surprises. The price you see is the price you pay, everywhere.

Thank you for being here

If any of this raises a question, about your plan, your credits, or what is right for your firm, please just reach out. A real person on our team will read it, and we would genuinely rather talk it through than have you wonder.

We know what it means to put your work in someone else’s hands. We do not take that lightly, and we intend to keep earning it, one faster answer, one cleaner draft, one honest invoice at a time.

You can see the new plans here, and if you would rather talk, you can always reach us directly.

With gratitude,
Abiola Ogodo
Founder, ModulawAI

This is a bigger update than usual, so it is worth going through properly. Four things changed: ModulawAI can now be driven by other AI assistants, there is a REST API for building your own integrations, the app has moved to a new address, and there is a full help centre where before there was very little.

Connect an AI assistant to your workspace

The headline change. ModulawAI now runs a Model Context Protocol server, which is the emerging standard for letting an AI assistant work with an outside system. In practice that means you can point Claude, or any other MCP-capable client, at your ModulawAI workspace and it will work with your real matters instead of guessing.

The difference is worth being concrete about. Without this, asking an assistant about a case means pasting context in by hand and hoping you included the right bits. With it, the assistant can look the matter up itself.

What it can reach

Twelve areas of your workspace are exposed:

• Cases and tasks, for looking up matters and the work outstanding on them
• Clients and intake, covering your roster and incoming enquiries
• Documents and notes, so an assistant can read what is on a file
• Calendar, for court dates and deadlines
• Financials, covering invoicing, expenses and time
• Forms and processes, for the structured work your firm runs on
• Workspace administration
• Legal research, which reaches the same case law corpus the in-app assistant uses

That last one is the one people tend to underestimate. It means an assistant working in your own tooling can search Nigerian case law directly, rather than falling back on whatever it happens to remember.

How the security works

Connecting an outside tool to a system holding privileged client material deserves scrutiny, so here is exactly how it is constrained.

• Standard OAuth, not a shared secret. Connections use OAuth 2.1. You approve a connection through a normal consent screen, and no credential is ever pasted into a third-party tool.
• Bound to one workspace. A token is issued against the workspace you approved it for and cannot reach into another.
• Read by default, write only if you say so. Write access is granted per credential, and you choose whether to grant it at all.
• Destructive actions are confirmed. Anything that deletes or irreversibly changes data requires an explicit confirmation rather than proceeding silently.
• Revocation actually revokes. Every connected app and agent is listed in your profile. Disconnecting one invalidates its access immediately.

Command-line clients that cannot open a browser are supported too, through the device authorization flow, so you get a short code to approve rather than being locked out.

A REST API for your own systems

If your firm runs its own software, or you want to automate something specific, there is now a REST API rather than an assistant sitting in the middle.

Create a key from Profile, then API access. Each key is:

• Scoped. Grant read, or read and write, per key. A reporting script never needs write access, so do not give it any.
• Optionally pinned to one workspace. Useful if you run several and want an integration confined to one.
• Rate limited individually. Limits apply per key, so a runaway script cannot degrade everyone else’s access.
• Independently revocable. Kill one key without disturbing the others.

One thing to know before you plan around it: programmatic access requires an active subscription and is not available during a trial. If you are evaluating ModulawAI and want to test an integration, talk to us and we will sort something out.

The API and the MCP connector share the same permission model underneath, so a scope means the same thing in both. There is no second set of rules to learn.

And ModulawAI can connect out to other tools too

The connector works in both directions, which is easy to miss. As well as letting an outside assistant into your workspace, ModulawAI can now connect out to other MCP servers and use their tools inside your chat.

Open the + button in the chat composer and you will find connections there, alongside Zoho, Google Calendar and your cloud storage. Add an MCP server by URL and its tools become available to the assistant while you work.

In practice that means if a service your firm relies on exposes an MCP server, whether that is a research provider, an internal system or something you have built yourself, the assistant can use it without us having to build a bespoke integration first. Both the modern streamable HTTP transport and the older SSE transport are supported, and credentials are encrypted at rest.

More models to choose from

Separately, the model picker has grown. Alongside the existing options you can now select:

• GLM 5.2, a frontier open-weight model with strong reasoning and a very long context window, useful when you are working across a lot of material at once
• Kimi K2.6, reliable at tool use and structured output, with strong reasoning. It takes slightly longer to produce its first words because it thinks before it answers
• MiniMax M3, frontier reasoning with a long context and native image understanding, so it is the one to reach for with scanned exhibits or photographed documents

These three sit on the Professional plan and above, and consume credits at twice the standard rate. That is a reflection of what they cost to run rather than a judgement about when you should use them, but it is worth knowing before you make one your default. The full picture, including which model suits which task, is in Choosing an AI Model and What Credits Cost.

The app has moved to app.modulaw.ai

ModulawAI now lives at app.modulaw.ai. The main modulaw.ai address becomes our public site, covering the product, pricing and the help centre.

Existing bookmarks continue to work and will send you to the right place, so nothing breaks if you do nothing. Still worth updating them when you get a moment. Your account, your workspaces and your data are entirely unaffected. Only the address changed.

A help centre worth reading

We have published a proper help centre, organised into collections rather than a single undifferentiated list. It currently covers:

• Getting started, including creating an account, signing in and a tour of the interface
• Legal Assistant, covering your first question, uploading documents, drafting, voice mode, choosing a model, and the limits of what the assistant can and cannot do
• Case and matter management, covering creating cases, notes, tasks and deadlines
• Workflows and automation, including what workflows can and cannot do
• Team chat and collaboration, covering channels, direct messages, and bringing the assistant into a channel
• Documents and connecting your email

Two of these are worth singling out. What the Assistant Can and Can’t Do is deliberately blunt about the limits, including what it will not do and where you must check its work. Understanding AI Answers and Citations explains how to read a citation, what the verification badge actually means, and why it means something slightly different depending on the jurisdiction. If you rely on this tool for research, read that one.

More guides are on the way, including the client portal and the CRM.

A rebuilt foundation

Less visible, but the reason several of the above were possible. We have moved ModulawAI’s infrastructure onto AWS: the servers running the application, the caching layer, document storage, document text recognition, and the AI models handling research and drafting.

You should not have noticed this happening, which was the goal. What you should notice is that responses come back faster and the platform holds up better when a lot of people are using it at once.

Getting started

If you want to try the assistant connection or the API, go to Profile, then API access in app.modulaw.ai. The connect instructions for MCP clients are on that page, along with a link to the developer documentation.

If you build something interesting with it, we would genuinely like to hear about it.

Ask any lawyer where the real work happens and the answer is Microsoft Word. Yet most legal AI lives in a separate browser tab, forcing a constant copy-paste shuffle that breaks concentration and loses context. We took a different route: we put ModulawAI inside the document.

Redlines through Word’s own track changes

Suggestions arrive as native track-changes edits, not as text to retype. Accept, reject, and review them exactly as you would a colleague’s markup, because that is what they are.

Findings as comments, anchored to the paragraph

When the assistant flags a risk or a missing clause, it writes a Word comment anchored to the exact paragraph. The finding lives where the issue is, with the reasoning attached.

Citations that hold up

Drafting and citations are jurisdiction-aware and checked as you write, so the authority you drop into a brief is one you can stand behind.

• Native track-changes redlines, right in the document.
• Findings delivered as Word comments you can act on.
• Jurisdiction-aware citations, verified as you draft.
• Every change tied to the matter and the audit trail.

Install it from Microsoft AppSource and the assistant is there the next time you open a document — no tab-switching, no lost context.