An AI agent can collect that data from a conversation, a file it can read, or other tools and integrations connected to the agent—such as a CRM, accounting system, or spreadsheet—and then pass it to ActiveMerge MCP to prepare a document for approval. ActiveMerge’s main job is to merge data into a document template and generate a finished file.
The distinction is useful: a source file can provide data for generation, while ActiveMerge’s Files library holds documents that already exist and can be listed or merged. ActiveMerge MCP does not upload arbitrary local attachments into that library.
How an agent uses data
An AI agent can gather data from any sources available to it, including connected business apps and tools. For example, it could retrieve customer details from a CRM, invoice terms from an accounting system, and line items from a spreadsheet, then combine those values for an ActiveMerge template. The agent’s available connections depend on the AI client and the integrations configured for that agent.
An AI client’s ability to read attachments also depends on that client. If it can read a purchase order, invoice request, timesheet, or project brief, the agent can extract the relevant values and map them to an ActiveMerge template. ActiveMerge MCP provides access to the user’s ActiveMerge templates and document-generation tools; it does not itself connect to the CRM or accounting system, nor read local attachments or import their contents.
The agent should discover the template’s required fields before preparing a request. It should use the template’s field names, ask the user about missing or ambiguous details, and avoid inventing customer, pricing, tax, or payment information.
ActiveMerge MCP capabilities
After you connect the MCP server, an AI client can use these capabilities:
| Capability | What the agent can do |
|---|---|
| Account and credits | Check the connected account and its available document credits. |
| Templates | List saved templates and retrieve the merge fields for a selected template. |
| Document generation | Submit template field values in data, and repeating rows in items, to prepare a DOCX, PPTX, or PDF. Generation only runs after you approve the exact request in ActiveMerge. |
| Generation status | Check a generation job and retrieve its download link when it is ready. |
| ActiveMerge Files | List files and folders already in the connected account’s Files library. These file IDs can be used as merge inputs. |
| Document merging | Combine 2–20 ActiveMerge files or public HTTPS document URLs into PDF, DOCX, or EPUB after approval of the exact inputs and options. DOCX and EPUB output require DOCX inputs. |
| Template drafts | Create a new AI-generated DOCX or PPTX draft for review. Draft creation costs 5 credits; it is saved only after you approve it. Existing templates are not changed. |
| Workflows | List workflows in the connected account. The MCP tools do not run a workflow. |
Example: prepare an invoice from order data
A client can ask an agent:
Create an invoice for Acme Ltd. from my ActiveMerge Invoice template. Use the attached order to fill in the customer details and service line items. Use invoice number INV-2048 and a due date of December 17, 2026. Show me the invoice details for approval before generating it.
The agent can then:
- Read the attached order if the AI client supports that file type.
- List the client’s ActiveMerge templates and select the Invoice template.
- Retrieve the template fields so it uses the correct names for customer details, invoice dates, totals, and other fixed values.
- Map fixed values into
dataand each service row intoitems. For example, a template with{{item.Description}},{{item.Qty}}, and{{item.Amount}}uses item rows like:{ "items": [ { "Description": "Consulting", "Qty": 2, "Amount": 500 }, { "Description": "Setup", "Qty": 1, "Amount": 150 } ] } - Summarize the invoice fields, amounts, and line items for the client. Ask about missing details, such as tax treatment or payment terms, instead of guessing.
- Request document-generation approval from ActiveMerge, then wait for the client to approve the request.
- Generate the invoice using the same template,
data,items, and output format that were approved. - Check the job status and return the download link when the invoice is ready.
The approval is bound to the exact template, field data, line-item rows, and output format. If any value or item changes, the agent must request approval again. The agent should verify invoice totals against the source order before asking for approval.
For line-item template setup and matching row names, see Generating documents with line items.
Other data-driven workflows
- Prepare a proposal from a project brief: Extract the scope, milestones, and pricing details, then map them to the saved proposal template.
- Create onboarding paperwork: Collect employee or customer details from a conversation or connected system, discover the needed template fields, and prepare the forms or letters.
- Generate a report after a support interaction: Turn a conversation summary and confirmed account details into a report or service confirmation.
- Create a reusable template: Describe the layout and fields for a new DOCX or PPTX draft, inspect it, then approve it to save as a new template.
When the ActiveMerge Files library is involved
The Files library is for documents already stored in ActiveMerge. MCP can list those files and use their IDs as inputs when merging documents. It can also merge from a public HTTPS document URL, subject to safety and size checks. It cannot upload a local attachment into the library.
This means an agent can use a source attachment to extract generation data, and separately use ActiveMerge file IDs to assemble existing documents into a packet. Those are different operations.
Connect ActiveMerge MCP
Add https://app.activemerge.com/mcp to an MCP-compatible AI client and authenticate with your ActiveMerge account. See How to connect ChatGPT and Claude to ActiveMerge for setup instructions and ActiveMerge MCP use cases for more examples.


