This release lets you connect Dive to your company systems on the Enterprise plan. The roster can be synced automatically from HR systems and Microsoft Entra ID, and the API (beta) lets company systems and AI tools search and read Dive SOPs. You can now register departments and positions in the roster, show each member's position on the Skill Map, and sort and filter by department and position. You can also trim the beginning and end of a screen recording before importing it. We also improved subtitle line breaks and how failed materials are shown, and fixed several issues with saving in the video editor.

On October 16, we released “AI connections (MCP),” which lets the AI tools you already use, such as Claude and ChatGPT, search and read Dive SOPs. Use AI connections for AI tools used by people, and use the new API for systems you have built in-house. For details on AI connections, see the 2026-10-16 release article.

①Dive can now connect to your company systems (Enterprise)

On the Enterprise plan, Dive can now connect to your company systems. There are two features: automatic roster sync from HR systems (SCIM), and an API that lets company systems and AI tools search and read Dive SOPs. Both are connected using a key issued by the owner.

The roster can now sync automatically from HR systems (SCIM)

Dive's roster can now be synced automatically from HR systems and identity management services such as Microsoft Entra ID and Okta. It supports SCIM 2.0, the industry standard. When someone joins, transfers, or leaves, Dive accounts are created, updated, or deactivated, and department, position, and employee number are updated automatically.

Until now, every time someone joined, transferred, or left, the roster in Dive had to be updated by hand on screen or in Excel. With this release, information from HR systems and identity management services flows into Dive directly, so you no longer need to edit the roster by hand.

●What sync can do

• Creates Dive users for people added in the HR system or identity management service. People with an email address are created with the status “Inviting,” and people without an email address are created as Email-free accounts
• Reflects changes to department, position, or name from transfers and the like in the Dive roster
• People who have left are also deactivated in Dive. Their viewing and training records are kept
• Under [Mapping] in [Department/Position], you can decide the group from the department in HR and the permissions from the position

●How to get started

The owner enables sync in the [HR roster sync (SCIM)] card on the [Security] tab under [Owner Functions] in the side menu, and issues a Key (secret token). Set the receiving URL and key shown on the card in your HR system or identity management service, and sync starts.

●Please note

• People created through sync count toward your Number of accounts. If creating them would exceed the Account limit, they are not created, and the owner is notified by email the next morning
• Invitation emails are not sent automatically to people created by sync who have the status “Inviting.” For teams using SAML SSO, the account is linked automatically when the person logs in with SSO for the first time. For other teams, go to [User management] and use Resend invitation email
• While sync is in use, the department, position, and HR items sent from HR, as well as the permissions and Group Affiliation determined by the mapping, can no longer be changed on the Dive screen
• While sync is in use, people who log in with SSO for the first time are not registered automatically

●Requirements

• Available on the Enterprise plan. The owner performs the setup

For details, see Automatically sync your roster from HR systems (SCIM).

Company systems and AI can now search for and read SOPs (API, beta)

You can now call the Dive API from your company systems to search for and read SOPs. For example, a chatbot you have built with generative AI can look up Dive SOPs when answering questions, company portal search can include Dive SOPs, and equipment or process screens can show the related SOPs.

On October 16, we released “AI connections (MCP),” which lets AI tools people use, such as Claude and ChatGPT, search and read Dive SOPs. AI connections read as the Dive account of the person who connected it, so they could not be used from systems built in-house. With this release, you can use an API key issued by the owner to search and read Dive SOPs from your company systems. People using these systems do not need a Dive account (you still need to log in to Dive to open videos).

●What the API can do

• Search: send a word, and it returns the matching published SOPs, with excerpts of the matching parts (including video subtitles) and a link to open them in Dive. It returns up to 20 results at a time
• Read: returns the SOP title, document number, overview, and items used, plus the work details, tips, quality, and safety for each step, and the body of File-based SOPs. Each step includes a link that opens that step in Dive
• You can only search and read. You cannot create, edit, or delete SOPs
• Only published versions are returned. Drafts are not returned. When a revision is published, the new version is returned
• The contents of videos, images, and Related files are not returned. Please view them in Dive using the links in the response

The API specification is also published in OpenAPI (3.0) format.

●Issuing keys and what they can show

The owner issues a key from [Issue key] in the [API (search and read SOPs)] card on the [Integrations] tab under [Owner Functions] in the side menu. For each key, you decide the following.

• Groups to show: the key returns only published SOPs within the range that View-only users belonging to all of the selected groups can see in Dive. SOPs in folders not restricted by group are always returned. If you do not select any group, only SOPs in folders not restricted by group are returned
• Source IP address: if you register an address, the key accepts calls only from that address
• Expiry: choose from 30 days, 90 days, or 1 year. As the expiry approaches, the owner is notified by email

A key is shown only once, when it is issued. You can test the key's scope by entering a word with [Test with this key] on the key's row. Each team can have up to 5 active keys.

●Checking how it is used

On the card, [Usage status] shows the number of API searches and reads for the last 7, 30, and 90 days, by key and by SOP. You can export the details, including the search words, using [Export details as CSV]. Issuing, changing, or revoking keys and exporting the details are also recorded in the audit log.

●Requirements

• Available on the Enterprise plan. Keys are issued and configured by the owner from a PC browser
• Teams that use IP address restriction must register the source IP address on each key
• Call the API from your company server. It cannot be called directly from the browser screen
• Provided as a beta version

Note: Before you use this feature, be sure to read “Compliance and Disclaimer” at the bottom of the how-to article. By using this feature, you are considered to have agreed to its contents.

For how to issue a key and how to call the API, see Searching and reading SOPs from company systems and AI (API, beta).

②You can now register departments and positions in the roster

On the roster in [User management], you can now register a department, position, and phonetic reading for each person. Departments can be organized in a hierarchy such as “HQ / Division / Section,” and in addition to the main department, you can register secondary departments.

Until now, the Dive roster only had groups and permissions, so company departments and positions had to be written into the display name, for example. With this release, departments and positions can be held as roster items. The roster list shows department and position columns, and you can filter by department and position. Searching the roster also works with phonetic readings.

●How to register

• The Owner or Team Administrator creates the list of departments and positions from [Department/Position] at the top of the [User management] screen. Positions are arranged so that higher positions are placed at the top
• Set each person's department, position, and phonetic reading in [Affiliation and position] under [User details]
• You can also register them in bulk with Excel from [Bulk change]. Phonetic reading, department, secondary department, and position columns (columns Q to T) have been added to the right end of the “User list” sheet, and “Department” and “Position” sheets have been added. Departments and positions not in the list are created during import

●Difference from groups and permissions

Departments and positions are used for display, sorting, and filtering in the roster and the Skill Map. Who can view which SOPs and approvals are still determined by groups and permissions, as before. Registering departments and positions does not change what people can view or their permissions.

●HR items

[User details] now has [HR items], where you can record employee number, hire date, employment type, and retirement date. Only Owners and Team Administrators can view them.

●Requirements

• Available on all plans
• Only the Owner and Team Administrators can create the list of departments and positions. The scope in which individual departments and positions can be changed is the same as for changing permissions (Group Administrators can only change users in the groups they manage)

For details, see Registering departments and positions (roster department, position, phonetic reading, and HR items).

③The Skill Map can now show positions and sort and filter by department and position

You can now show each member's position in the member headings on the Skill Map. You can also sort by position or department, and narrow down the people displayed by department and position.

Until now, member headings on the Skill Map showed permissions (such as Owner and Group Administrator). Because skill matrices on the floor are often viewed by position, this release shows positions for teams that have registered positions.

●What you can do with [Order & display]

• Heading tag: choose from [Position], [Permission], or [Hide]. Teams with at least one position registered show positions by default. Guests are still shown as “Guest,” as before
• Order: [Position: top first], [Position: bottom first], and [Department order] have been added. For the [Name] order, people with a phonetic reading are sorted by that reading
• Filter: use [Filter by department/position] to narrow down the people displayed by department and position. Selecting a parent department also includes people in its sub-departments. Goal and achievement counts are calculated across all group members before filtering

Positions are also shown on personal pages, in the list of trainees for each training manager, and in the skill level status for each skill. In the Skill Map Excel export, department and position columns are added at the right end of the “Evaluation details” and “Members” sheets.

●Requirements

• Available on the Core, Pro, and Enterprise plans (Excel export of the Skill Map is available on the Pro and Enterprise plans)

④You can now trim unneeded parts of a screen recording before importing it

In the [Recording preview] that appears after you stop a screen recording, you can now set the range to import. You can trim the unneeded parts at the beginning and end of the recording before importing.

Until now, the whole screen recording was imported, so unneeded parts at the start of the recording or just before you stopped had to be cut after import. With this release, drag the handles at both ends of [Range to use] to set the range, then click [Continue with this range], and only that range is imported as material. AI analysis is also performed on the video within the range you set.

If you click [Continue] without changing the range, the entire recording is imported as before.

⑤Other improvements and bug fixes

●Subtitles

• When subtitles are shown on two lines, you can now choose where the line breaks, based on the end of a sentence, punctuation, spaces, or word boundaries. Previously, the two lines were made the same length first, so a line could break in the middle of a word (for example, “...after removing 4 | kara, ...”). For Japanese text, choose a break that does not fall before a particle (for example, “bolt | wo”) or inside a word (for example, “air | filter” or “safety | check”)
• On smartphones in portrait view, the subtitle background bar is no longer fixed at 90% of the screen width. It is now sized to fit the subtitle, so short subtitles appear in a short bar
• When the text panel on the right is closed on a PC, or when viewing on a large screen, a single subtitle line is prevented from becoming too long
• Subtitles shown over the video in the video editor and in MP4 export are reflected in the same way

●Materials

• When you open a video that failed to process from the material details screen, the screen now shows “Processing failed” and the reason for the failure. You can also reprocess it from that screen with [Reprocess]
• For failures that reprocessing cannot fix (for example, the file is corrupted, has no video, is an image file rather than a video, or exceeds the length limit), “Re-upload recommended” is shown instead of the reprocess button
• Files whose contents are images rather than video (such as preview images saved with a video file name), and videos shorter than 1 second, are now marked “Processing failed” at the import stage, with the reason shown. Previously, the import continued and the failure might occur during AI analysis
• Videos that failed to process are no longer shown as “AI processing”
• Videos that failed during the conversion stage are now also reported by email to the person who uploaded them

●Video editor and SOP editing

• Fixed an issue in SOPs with pasted steps, where changes made in the video editor were not reflected or subtitles reverted to their original state
• Fixed an issue where, when you applied suggestions from AI proofreading or the Pre-publish check using [Apply], the text in video-based steps could revert later, after you finalized in the video editor or reopened the SOP
• Fixed an issue where edits made while saving were sometimes not saved, and an issue where “Save completed” was sometimes shown even when saving failed. When saving fails, “Save failed” is shown, and if there are unsaved changes, the [Save] button is shown again

●MP4 export

• Improved an issue where, during busy periods, MP4 creation failed and a notification email was sent. During busy periods, creation waits for a free slot and then continues

●SOP playback screen

• Fixed an issue on PCs where, when the window was narrowed, the [Correct misreading] ([Report misreading]) button at the top of the screen overlapped with other buttons. At widths where they would overlap, only the icon is shown

Impact on existing customers

The contents and settings of SOPs have not changed. Please check the following points.

• If you do not register departments or positions, the display of the roster and Skill Map is the same as before. Once you register at least one position, the Skill Map headings switch to showing positions (you can switch back to showing permissions from [Order & display])
• Columns have been added to the “User list” sheet of the [Bulk change] Excel file. Files you downloaded previously can also be imported as they are (in that case, phonetic reading, department, and position do not change)
• HR roster sync does not run until the owner enables it
• The API is not used until the owner issues a key. Below the [AI connections (MCP)] card on the [Integrations] tab under [Owner Functions], we have added the [API (search and read SOPs)] card. The card is also displayed on plans other than Enterprise, but keys cannot be issued on those plans
• Subtitle line-break positions and bar widths are reflected in playback screens and in MP4 files exported from now on. Exported MP4 files already created do not change. To use the new display, export again

Thank you for your continued support of Dive◎