Legistify Codex: Complete User Guide

Legistify Codex is an AI-powered workspace built into the platform that lets you draft documents, analyse files, run research, and make structured decisions, all without switching between tools. Whether you are working on cases, contracts, intellectual property, or notices, Codex is available across all four modules: Litigation Management (LMT), Contract Management (CMT), Intellectual Property Rights (IPR), and Notice Management (NMT).

This guide covers everything you need to know to get started and make the most of Codex, from writing your first prompt to managing multi-item projects.

Codex is an AI workspace embedded within the Legistify platform. It works like a conversational AI interface where you type in plain language and receive structured, ready-to-use outputs.

You can use Codex to:

Codex is designed to support your workflow, not replace your judgement. It gives you a structured starting point that you can review, refine, and act on.

Codex is available in all four modules of the Legistify platform. The steps to access it are the same across all modules.

When you open Codex, you will see a clean workspace with a prompt bar at the centre of the screen. Here is a quick overview of what each element does so you know your way around before you start.

1. The prompt field
This is where you type your questions, instructions, or tasks. It is labelled "Enter a prompt for Codex." This is the main way you interact with Codex.

2. The + icon
Located on the left side of the prompt bar, the + icon is your main menu for all Codex actions. Clicking it opens a menu with four options: Upload Files, Create Project, Select Project, and Use Style. Whenever this guide refers to any of these features, you will always find them here.

3. The hamburger menu (three lines icon)
Located to the right of the + icon in the prompt bar. Clicking it opens a side panel showing New Chat, Projects, and Recent Chats. This is where you manage your conversations and projects.

4. New Chat
Starts a fresh conversation with no prior context. Use this when you want to begin a completely new task that is unrelated to your current conversation.

5. The model selector
Displayed in the bottom right of the prompt bar, this shows the currently active AI model. Click on it to switch to a different model depending on your task.

6. The send button
The upward arrow on the far right of the prompt bar. Click this to submit your prompt.

7. Predefined category buttons
On the Codex home screen, below the prompt bar, you will see a set of category buttons. These are predefined prompt suggestions grouped by task type to help you get started quickly. They differ depending on which module you are in.

8. Create Project and All Projects buttons
Also visible on the home screen below the category buttons. Use these to create a new project or view all your existing projects.

Prompting is how you communicate with Codex. You simply type your question, instruction, or task into the input field and Codex generates a structured response.

Type your requirement clearly in the prompt field that reads "Enter a prompt for Codex." The more specific and detailed your prompt is, the more relevant and structured the output will be.

For example, instead of typing "draft something," try "draft a termination clause for a SaaS agreement that protects the service provider." The added context helps Codex produce a more useful and accurate response.

Once you have entered your prompt, click the send button (the upward arrow) to submit it.

Codex will process your prompt and generate a response directly in the chat. Depending on your request, the response may include drafted text, summaries, bullet-point breakdowns, risk assessments, checklists, or structured explanations. You can copy the response, ask follow-up questions, or refine your request within the same conversation.

For drafting tasks, you may also see an Open button alongside the response. This opens the draft as a fully formatted, editable document in the Artifacts workspace. Read more about this in the Artifacts section below.

To help you get started, Codex provides predefined prompt suggestions on the home screen, grouped by category. These are suggestions you can click on to quickly run a common task. You are not required to use them. You can always type whatever you need directly into the prompt field.

The categories shown differ depending on which module you are in:

Litigation Management Tool (LMT)

Contract Management Tool (CMT)

Intellectual Property Rights (IPR)

Notice Management Tool (NMT)

You do not need to start a new chat for every task. After Codex generates a response, you can ask follow-up questions or refine your request within the same conversation. Codex retains the context of the current chat, so your follow-up prompts build on what was already discussed.

Codex allows you to attach documents directly within the workspace. This is useful when you want Codex to analyse a specific file rather than work from a general prompt alone.

Codex will analyse both your prompt and the document content before generating a response. This means the output is based on the actual content of your file, not a generic answer.

Use Style lets you control the tone and format of Codex's responses. Instead of adjusting your prompt each time, you can set a style preference that applies to the current session.

Choose the style that best fits the task you are working on. You can change the style at any point during your session.

Codex gives you the flexibility to choose from multiple AI models depending on the type of task you are working on. Different models have different strengths, so selecting the right one can improve the quality of your output.

The currently selected model is displayed in the bottom right of the prompt bar. Click on it to open the model selector and choose a different option.

1. Claude Opus 4.7
The most capable model for legal work. Best suited for complex tasks that require deep reasoning, nuanced analysis, and high-quality drafting.

2. Claude Opus 4.6
Specialised in legal drafting and processing. A strong choice for drafting agreements, clauses, and structured legal documents.

3. Claude Haiku 4.5
Specialised in contract analysis. Well suited for tasks that involve reviewing, extracting, and summarising contract content quickly.

4. More Models (under the "More models" option)

As a general guide:

You can switch models at any point. If a response does not meet your expectations, try the same prompt with a different model.

Projects is one of the most powerful features in Codex. It allows you to group multiple items (cases, contracts, or IPs) into a single workspace and run queries across all of them together. Instead of analysing one item at a time, you can work at a portfolio level.

A Project is a dedicated workspace within Codex where you bring together a set of related items and work on them collectively with shared context. You define the project name, a description, and custom instructions that tell Codex how to respond. Every query you run inside the project applies those instructions automatically and considers all selected items at once.

When managing multiple items, it is common to:

Projects solve this by:

For example, you can group all high-value Mumbai cases into a project, set an instruction like "keep responses concise and focus on financial risk," and then ask Codex to identify missing information across all of them in a single query.

The steps to create a project are the same across all modules. The only difference is what you select in step 5.

​1. Go to Codex in your module.

​2. Click Create Project (visible on the Codex home screen or via the + icon in the prompt bar).

3. A side panel will open. Fill in:

4. Click Next.

5. Select the items you want to include in the project:

6. Click Create Project.

Once your project is created, it is automatically set as the active project. You will see the project name tagged in the prompt bar, confirming it is active. Simply start entering your queries and Codex will apply your instructions and context across all selected items automatically.

If you want to switch to a different project:

All responses will automatically follow your defined instructions and consider all selected items in the active project.

If you want to exit a project and go back to a regular Codex chat, click the x next to the project name in the prompt bar. This deselects the project and Codex will return to responding without any project context.

You can access all your projects in two ways:

Custom instructions are what make Projects powerful. Here are some tips to get the best results:

Good custom instructions mean you get consistent, predictable outputs every time without needing to repeat yourself in every prompt.

Artifacts is a workspace feature within Codex that turns AI-generated drafts into formatted, editable documents. Instead of receiving a response as plain text that you then have to copy into another tool, Artifacts opens the draft as a structured document right inside Codex.

Artifacts are triggered specifically by drafting tasks. When you ask Codex to draft a document such as a contract, clause, notice, or agreement, the response will include an Open button. Clicking it opens the draft as a fully formatted document in the Artifact panel alongside the chat. From there, you can view it, edit it manually, or ask Codex to refine it further, all without leaving the platform.

Note that Artifacts do not appear for non-drafting tasks such as summaries, risk analysis, or general questions. They are designed specifically for document creation and refinement.

Typical drafting workflows involve multiple steps: generating content, copying it into an editor, making changes, going back to AI for updates, and repeating. This creates unnecessary friction.

Artifacts eliminates that gap by:

1. Go to Codex in your module.

2. Enter your drafting requirement in the prompt field. For example: "Draft a mutual non-disclosure agreement suitable for sharing confidential business information."

3. Submit the prompt.

4. Once Codex generates the draft, you will see an Open button alongside the response.

5. Click Open to view the draft in the Artifact panel (a side workspace that opens to the right).

6. Inside the Artifact, you have two modes:

7. To refine the document using AI, type your update request in the prompt bar (for example: "Make the indemnity clause more balanced" or "Add a clause on data protection").

8. Codex will generate a new version of the document with your requested changes applied.

9. You can open any previous version at any time to compare or revert.

10. Once you are satisfied with the document, you can:

1. In-platform editing
Edit your document directly within Codex without exporting it to another tool.

2. View and Edit modes
Switch between a clean formatted view for reading and an edit mode for making changes.

3. AI-powered refinement
Use the prompt bar to ask Codex to modify specific sections, change the tone, add or remove clauses, or restructure the document.

4. Version control
Every AI update creates a new version. You can go back to any earlier version at any point, giving you full control over the drafting process.

Artifacts are most useful when:

Every conversation you have in Codex is automatically saved. To access your chat history, click the hamburger menu (three lines icon) in the prompt bar. A panel will open showing three options. New Chat starts a fresh conversation with no prior context, useful when you want to begin a completely new task. Projects takes you to your list of projects. Recent Chats shows all your saved conversations, each listed by the prompt you entered. If a chat was run within a project, the project name is tagged alongside it so you can easily identify it. Click on any chat to reopen it and continue where you left off.

Getting the most out of Codex comes down to how you interact with it. Here are some practical tips:

1. Be specific in your prompts

The more context you give, the better the output. Instead of "summarise this contract," try "summarise the key commercial terms, payment obligations, and termination conditions in this contract." Specific prompts lead to structured, relevant responses.

2. Give Codex a role or perspective

You can tell Codex whose perspective to take. For example: "Review this agreement from the buyer's perspective and identify any unfavourable clauses." This helps Codex tailor its analysis to what actually matters to you.

3. Use follow-up questions

You do not need to write one perfect prompt. Start broad and then narrow down. Ask a follow-up like "Now focus only on the indemnity clause" or "Can you simplify this in plain English?" Codex remembers the context of your current conversation.

4. Try a different model if the output is not right

Different models have different strengths. If a response feels too generic or lacks depth, switch to a more capable model like Claude Opus 4.7 and try the same prompt again.

5. Use Style to control the format

If you find yourself repeatedly asking for shorter responses or more formal language, set a style preference using the + icon. It saves time and keeps outputs consistent throughout your session.

6. Use Projects for recurring workflows

If you regularly work with the same set of cases or contracts, create a Project. Set your instructions once and you will not need to repeat context every time you open Codex.

7. Upload the actual document

When your task involves a specific file, always upload it rather than describing it. Codex analyses the actual content and gives you far more accurate and relevant outputs.

This guide covers all current features of Legistify Codex. As new capabilities are added, this article will be updated accordingly.