MemoMeister
DocumentsConnect AI agents to MemoMeister, a construction and trades documentation platform from Freiraum GmbH. Agents file field memos into project folders, label them, control folder permissions, and search across active memos with autocomplete.
What This Integration Enables
MemoMeister is documentation for construction and trades, and the thing worth being clear about is what a memo is for. It is not a file store. Site photos, delivery notes, defect reports, and signed PDFs filed with the right date in the right project folder are the record of what the site looked like on a given day, and that record is what gets produced when somebody disputes what happened. Automation that files well is worth a great deal here. Automation that files carelessly creates an archive nobody trusts, which is worse than no archive at all.
The connector covers the whole shape of that job. The folder tree is fully manageable: Create Folder can build an entire project structure in one call from a folder template, Update Folder pins the site address on the map after resolving it through Find Locations By Address, and Archive Folder retires a job without deleting its documentation. Memos are created as text or from a file at a publicly reachable URL, downloaded back out into FlowRunner file storage in any of five stored renditions, moved singly or in batches, copied, labelled, pinned, commented on with mentions, and either binned recoverably or deleted permanently. Labels are MemoMeister's cross folder tagging mechanism, so a flow can tag site photos consistently regardless of which project they land in, and a label that does not yet exist is created automatically when it is applied.
Several behaviours reward knowing in advance. MemoMeister is a GraphQL API, so a failed query commonly arrives as HTTP 200 with an errors array rather than an error status. This connector inspects every response and raises a step failure when errors are present, so a flow never silently continues on a rejected query. Separately, several mutations carry a userErrors array for per-item problems, returned to the flow rather than raised, because on batch operations such as Move Memos To Folder some items succeed while others fail. Check userErrors on any batch result before treating it as done. A memo's location can only be set at creation, since the API exposes no mutation to change it afterwards, which makes the capture step the only chance to get the pin right. And the API key inherits every permission of the user who created it, with no scopes and no read-only mode, so the narrowest sensible integration user is a real design decision rather than a formality.
There are no triggers. MemoMeister exposes GraphQL subscriptions over a websocket, which FlowRunner triggers cannot consume, and publishes no outbound webhook registration endpoint. Poll instead: List Memos with Modified After and a descending sort on last modified for new or changed memos, List Recent Comments for new comments, and List Folders with Created After or a last activity sort for folder changes.
Without FlowRunner
With FlowRunner
Use Case Scenarios
A won job builds its own documentation structure
A job is marked won in the estimating system. The flow calls Create Folder with the folder template identifier for that job type, which builds the whole subfolder structure in a single call rather than a dozen. It then calls Find Locations By Address to resolve the site address into coordinates and Update Folder to pin it, supplying explicit coordinates alongside the address because the geocoding is OpenStreetMap based and some addresses resolve badly. Finally it calls Add Folder Permissions for the project manager and the assigned crew group. The identifiers returned are stored against the job record in Google Sheets, because MemoMeister identifiers are base64 global identifiers that must be passed back exactly as returned, and keeping a local mapping is far more reliable than searching by name on every future run.
Photos that arrive already classified
Site photos land in a shared folder in Google Drive or arrive by email. The flow calls Create File Memo with the file URL and the target project folder, letting MemoMeister index the content and generate previews. Then it does the part a person on a ladder will never do reliably: it runs the image through AWS Rekognition, and where a finding comes back, it writes the result with Create Memo Comment mentioning the site manager and applies a classification with Add Labels To Memo. Because labels work across folders, a defect tagged on one project is findable alongside every other defect in the company. If the photo carries EXIF location data MemoMeister uses it, so a photo taken on site pins itself, and a photo forwarded from an office does not.
Closing a job without losing the record
When a job completes, the flow archives rather than deletes. Archive Folder keeps the documentation intact while removing it from everyday search, and Restore Folder From Archive Or Bin brings it back if a warranty claim reopens the job. Where cleanup genuinely is required, the flow uses Move Items To Bin, which accepts folder and memo identifiers mixed in one list and is recoverable, rather than Delete Folder, which is permanent and takes the folder's contents with it. That preference is written into the flow deliberately: a construction archive is a legal record, and the recoverable option should be the default even when it is the slower one. A summary of the archived project goes to the operations channel in Slack with the folder link attached.
Human-in-Loop Highlight
The gate here is not on a deletion, which is the obvious candidate, but on Update Subfolder Permissions And Labels. Three facts about this API combine into something genuinely dangerous. Permissions in MemoMeister are per folder and are never inherited, so creating a subfolder does not copy the parent's access and the interface merely prefills it. This action is therefore the only practical way to apply an access set across an existing subtree. And it offers two modes: Add merges the set into what is already there, Replace overwrites it. Replacing a single individual permission is not supported by the API, so Replace is all or nothing across everything it touches. A flow written to onboard a new subcontractor and passing Replace instead of Add does not fail, does not warn, and does not partially apply. It quietly removes every other person's access to an entire project's documentation, and the way the company finds out is a foreman standing on a roof unable to open the drawings. Owners and admins still see folder names, which makes it look less broken than it is. So the agent assembles the change, reads the current state with Get Folder and List Users, resolves the group membership through Get Group, and stops: "Applying access for Kellner Dachbau to Projekt Nordhalle and 11 subfolders. Mode: Replace. This removes 6 existing permissions across those folders, listed below. Confirm Replace, or switch to Add." A site manager answers, and the mutation runs on that answer. This is human-in-the-loop applied to blast radius rather than to destruction: nothing here deletes a file, and that is exactly why it would otherwise slip through unreviewed.
Agent Capabilities
54 actionsAccount
4- Get My User Returns the user the API key acts as. Worth calling once in a flow's setup, because the key inherits that user's permissions entirely and everything the flow can reach follows from it.
- Get Active Company Returns the active company context for the key.
- List Users Lists the users in the company. The basis of any access review and of resolving a name to the identifier a permission call needs.
- Get User Retrieves one user's profile.
Folders
10- Get Folder Retrieves one folder with its metadata and section, meaning whether it sits in the normal tree, the archive, or the paper bin. The read to run before any permission change.
- Find Folder By Name Finds a folder by name. Useful for reconciliation, though the vendor's own guidance is to keep a local mapping from your record identifier to the MemoMeister identifier rather than searching by name on every run.
- List Folders Lists folders, with three distinct behaviours: omitting the parent returns folders at any level the user can see, which is the densest way to mirror a whole tree, Root Level Only returns just the first level, and supplying a parent returns that folder's direct subfolders.
- Create Folder Creates a folder, optionally building an entire subfolder structure from a folder template in one call. Folder names must be unique within one parent.
- Update Folder Updates a folder's details, including its address and map pin, and exposes a Clear Location option that removes a location cleanly rather than leaving it half set.
- Move Folder Moves a folder to a new parent. Fails when the target is the folder itself or one of its own descendants, or when the key's user cannot see one of the two identifiers.
- Archive Folder Moves a folder into the archive, keeping the documentation intact while removing it from everyday search. The right action for a closed job.
- Restore Folder From Archive Or Bin Brings a folder back from either the archive or the paper bin. The reason a flow should prefer archiving and binning over deleting.
- Move Items To Bin Moves folders and memos to the paper bin, accepting both kinds of identifier mixed in one list. Recoverable, and the sane default for any cleanup step.
- Delete Folder Permanently deletes a folder and everything inside it. Not recoverable, and it takes the folder's contents with it, so it should never be the default cleanup action in an automated flow.
Permissions
4- Add Folder Permissions Grants users or groups access to a folder. The additive, narrow operation, and the one most onboarding flows should use.
- Remove Folder Permissions Revokes access to a folder for named users or groups.
- Update Subfolder Permissions And Labels Applies a permission and label set across an existing subtree, in either Add mode to merge or Replace mode to overwrite. Because permissions are never inherited this is the only practical way to change access at scale, and because replacing one individual permission is not supported, Replace overwrites everything it reaches. The action this page argues should always pass through a person.
- List Folder Templates Lists the folder templates available, which is how a flow finds the structure to build a new project from.
Memos
19- Get Memo Retrieves one memo with its metadata, labels, and attachment details.
- List Memos Lists memos with filters including Modified After, media type, and sorting. Capped at 250 memos per request by the API, and the main polling surface in place of triggers.
- Create Text Memo Creates a text memo in Markdown, supporting headings, bold, italic, strikethrough, lists, todo lists, and links, though not tables or inline images. Every text memo also gets a generated PDF version.
- Create File Memo Creates a memo from a publicly reachable file URL, which the connector fetches and uploads. Images, PDFs, Office documents, email files, and videos receive processing including full text indexing and preview generation. A memo's location can only be set here, at creation.
- Replace Memo File Adds a new revision holding a new file. Despite the name it does not overwrite history, which the vendor documents for PDF, Word, and Excel memos.
- Download Memo File Downloads one of five stored renditions, meaning original, PDF preview, printable PDF, video preview, or audio preview, into FlowRunner file storage. Text memos have no original file, so use the PDF preview for those.
- Update Memo Updates a memo's metadata. Every metadata change creates a new revision, so the history stays intact.
- Update Text Memo Content Updates the Markdown content of a text memo.
- Set Memo Labels Replaces a memo's label set with the one supplied.
- Add Labels To Memo Adds labels to a memo without disturbing the ones already there. The additive counterpart to Set Memo Labels, and usually the one an agent should reach for.
- Remove Labels From Memo Removes named labels from a memo.
- Move Memo To Folder Moves one memo to another folder. Moving a memo between folders does not create a revision.
- Move Memos To Folder Moves several memos at once, which is the efficient batch option. The vendor documents that its document field is always null, so read userErrors to find out what actually happened.
- Copy Memo Copies a memo, used when the same evidence genuinely belongs in two project records.
- Rotate Memo Image Rotates a memo's image. Small, and useful in a flow that files photos captured on a phone held sideways.
- Set Memo Pinned Pins or unpins a memo in its folder, which is how a flow surfaces the item a crew needs to see first.
- Move Memo To Bin Moves a memo to the paper bin. Recoverable, unlike Delete Memo.
- Restore Memo From Bin Restores a memo from the bin, and requires an explicit target folder rather than assuming the original one.
- Delete Memo Permanently deletes a memo including its revisions. Not recoverable, so a retention flow should bin first and delete only on a separate, deliberate pass.
Comments
4- Create Memo Comment Adds a comment to a memo, with support for mentioning a user directly. The cleanest way for an agent to hand a finding to a named person inside the record it concerns.
- Update Memo Comment Edits an existing comment.
- Delete Memo Comment Removes a comment.
- List Recent Comments Returns the newest comments across every memo the key can see. The polling surface for new discussion, since there are no triggers.
Labels
5- List Company Labels Lists the labels defined for the company. Labels are cross folder, which is what makes consistent tagging across projects possible.
- Get Label Retrieves one label.
- Create Label Creates a label with a description. Only needed when the label should carry one, since applying a label that does not yet exist creates it automatically.
- Update Label Updates a label's name or description.
- Delete Label Deletes a label, which removes it from every folder and memo carrying it across the whole company. A small call with company-wide reach.
Groups
5- List Groups Lists the user groups in the company.
- Get Group Retrieves one group with its membership. The read that turns a group into a concrete list of people before an access change is proposed.
- Create Group Creates a user group with an initial set of users.
- Update Group Updates a group's name or membership. Because folder permissions are commonly granted to groups, this changes access everywhere the group is used.
- Delete Group Deletes a group along with the access it conferred.
Search
2- Search Autocomplete Runs MemoMeister's autocomplete search across the content the key can see. Archived and binned items are excluded unless a query asks for them explicitly.
- Find Locations By Address Resolves an address to coordinates through OpenStreetMap. Some addresses resolve inaccurately or not at all, so supply explicit coordinates alongside the address when the pin needs to be right.
Advanced
1- Run GraphQL Operation Runs a raw GraphQL operation against the MemoMeister schema. The escape hatch for anything the schema exposes that is not modelled as a dedicated action, including export jobs, dashboard widgets, report templates, and revision history.
Frequently Asked Questions
What can FlowRunner do with MemoMeister?
FlowRunner agents can run Get My User, Get Active Company, and List Users in MemoMeister, plus 51 more actions.
Does connecting MemoMeister to FlowRunner require OAuth?
No. MemoMeister connects to FlowRunner with an API key, no OAuth flow required.
Can MemoMeister trigger a FlowRunner workflow automatically?
MemoMeister doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with MemoMeister
$100 in credits. No card required. Connect in minutes.