HotDocs MCP. Automate complex document assembly from natural language prompts.

You connect your AI agent directly to document assembly workflows using HotDocs Advance. It lets you manage complex contract creation and reporting without ever clicking through a web interface. Your agent handles the whole process—from figuring out what documents exist, collecting all necessary data, generating the final file, and even auditing incomplete submissions.

Your agent first uses list_template_packages to check every template package available in your account, letting you determine exactly what kind of document structures can be assembled. You can then run list_template_versions on a specific package ID to see its full version history. If you need to find out the structural details of a particular template package, calling get_template_package gives you that info before you start.

Once you know what you're working with, your agent needs a reference point. It uses create_work_item to establish a unique record in the system, linking everything to a specific template package and ID. You can find existing records using list_work_items, which lets you filter by user IDs or date ranges, or narrow it down even more by calling list_work_items_by_date with an ISO 8601 start and end date format.

If you need to pull all the raw data associated with one of those work items at any point in time, get_work_item fetches that complete record.

When it comes to gathering information for a document, your agent has two main options. It can kick off an interactive session using create_interview_session, which guides your client through answering the specific variables needed for the template. If you already have all the data—say, if it's coming from another system—you don't need to talk to anyone; instead, you inject bulk answers into a work item by calling update_answers with a structured XML payload.

If there’s any chance that variables were missed during collection, your agent doesn't guess. It checks for gaps using get_unanswered_variables, listing every required field from the template that never got an answer. Once you confirm all data is ready and complete, calling complete_assembly generates the final documents—PDFs or Word docs—using all the collected data for a specific work item version.

The generated files are managed by other tools. You can see every document created for a particular work item by running list_documents, which gives you their file IDs and metadata. To actually get the content of that finished file, your agent uses get_document_content with the unique ID after assembly is complete.

If you need to check what documents were generated before calling list_documents, you can use it on a specific work item ID.

For troubleshooting or auditing purposes, there are other functions. You can retrieve an up-to-date API access token using get_auth_token if any of the other tools fail with authentication errors. Additionally, when generating documents, calling list_documents gives you the metadata and file IDs for all generated files associated with a given work item.