Welcome to Volume 9 of our 2026 product release notes! We've been working hard to bring you performance improvements that add value to your work. Read on to learn about the product enhancements and bug fixes released in October 2026.
Benchling AI
Benchling's artificial intelligence agents and features.
Public directory for AI Connectors and Models [2026.9.AI.1]
On https://www.benchling.com/directory we have a new public directory that lists available connectors that can be added to Benchling Chat.
File-created event triggers for Agents [2026.9.AI.2]
Agent can now be triggered when a File is created.
Notebook Check fixes default to full-actions mode [2026.9.AI.3]
When a user chooses Fix all or Fix selected on Notebook Check suggestions, the resulting Benchling AI chat now opens in full-actions mode by default. Edits apply immediately without an extra approval step.
Agents can ask users structured follow-up questions [2026.9.AI.4]
Benchling agents now have an AskUser tool that lets them pose structured questions (such as multiple-choice or form-style prompts) to the user mid-run, replacing free-text prose for cases where a specific user input would be better.
Benchling AI chats show unread indicators [2026.9.AI.5]
Benchling AI chat history now shows a blue dot next to chats that have unread responses, so users can quickly spot and open chats they have not yet reviewed. Unread indicators appear in both the collapsed and expanded chat panels for Automated and Recent chats.
Benchling Slack app uses Slack file attachments as context [2026.9.AI.6]
Files attached to a Slack message that invokes the Benchling Slack app are now included as context for the assistant alongside the message text.
Voice now works in Default-mode [2026.9.AI.7]
Voice can now run for users with only access to Default mode. In this mode, Voice does not have access to Full actions tools.
Voice chats now get auto-generated titles. [2026.9.AI.8]
Voice chats in Benchling AI now automatically generate titles from their transcripts.
BoltzGen available in Pipelines at 1000 predictions per job [2026.9.AI.9]
When using the new BoltzGen Analysis pipeline step, inputs can produce up to 1000 generated sequences.
Run Scientific Models Pipeline Step [2026.9.AI.10]
Analysis pipelines can now create a "Run Scientific Model" step which allows inputting dataframes to run structure prediction and property prediction on sequences and SMILES.
DeepViscosity antibody viscosity model in Model Hub [2026.9.AI.11]
Model Hub now offers DeepViscosity, an antibody viscosity prediction model. Users can select DeepViscosity from Model Hub to predict viscosity for antibody sequences.
Agents list keeps scroll position when toggling an agent [2026.9.AI.12]
Pausing or resuming an agent from the Agents tab of the AI settings modal now keeps the list static, so the scroll position no longer resets.
Agent run names mirror chat session names [2026.9.AI.13]
Agent runs in the overview list now display the chat name instead of the agent name when a chat has been renamed.
Admins can manage Agents from the Tenant Admin Console within a single section [2026.9.AI.14]
Tenant admins can now manage Agents directly from the Tenant Admin Console with their own permissions. Settings for chat, agent guidelines, agent actions, and tool permissions are now grouped under a single Agents section.
Direct message with Agents [2026.9.AI.15]
Agents now support direct messages so users can chat with them as needed.
Agent Foundries can be renamed via chat [2026.9.AI.16]
Coordinator agents can now update foundry names.
Create multiple sequences in one chat [2026.9.AI.17]
DNA and AA sequences can be created in bulk via chat.
Voice responses are richer [2026.9.AI.18]
Voice responses now show entity chips and spoken replies are shorter. Refused actions are also better explained.
Voice sessions use Full actions for hands-free data capture [2026.9.AI.19]
Users with Full actions can now start a Voice session in a single click.
Voice sessions share history with text chat [2026.9.AI.20]
Completed Voice conversation transcripts and tool calls are now included as context in later text-chat sessions and in text compaction
Use chat for Data Import [2026.9.AI.21]
Data import mode allows users to easily upload files and structure them in Benchling using an out-of-the-box skill.
Property prediction models allow batches of up to 1000 jobs [2026.9.AI.22]
Property prediction models support larger batch sizes now, and are able to run up to 1000 prediction jobs in a single batch.
Agents can configure scientific model pipeline steps [2026.9.AI.23]
Benchling AI can now add and configure scientific model steps in pipelines to run property or structure predictions.
Updated copy on per-user credit limit feature to clarify functionality [2026.9.AI.24]
Updated the Credit limits description paragraph in AI Usage page to clarify that monthly limits apply across users, teams, and organizations, and that users are blocked from credit-required features until the next billing period once the limit is reached.
Canvas artifacts are @-mentionable in Notebook entries [2026.9.AI.25]
Vibe coded Canvas Artifacts can now be referenced in Notebook entries via @-mentions.
[Open LA] Science Foundries are now in Preview [2026.9.AI.26]
Science Foundries can be used to run scientific processes end-to-end with agents and scientists working as a team. Find out more information here.
[Open LA] Agents can now be run with their own permissions [2026.9.AI.27]
Agents can now be run with their own permissions and shared with users on the tenant. More information can be found in our Help Center here.
Documents (includes Notebook)
Benchling Notebook helps users log their experiments, create protocols, and capture results against their samples.
Template publishing now includes review and approvals workflow [2026.9.DOC.1]
Review processes can be applied to template collections, preventing publishing until a template or subtemplate has been approved. Template authors or reviewers can create an entry from a draft template to confirm changes.
To enable this feature contact Benchling Support
New excerpts and previews default to full width [2026.9.DOC.2]
A tenant configuration now makes new excerpt and preview blocks in notebook entries default to full width, so entries no longer require manual per-excerpt resizing to fit the full page width.
To enable this feature contact Benchling Support
Descriptive S3 paths now consistent across PDF and HTML backups [2026.9.DOC.3]
Human-readable backups now use the same descriptive source and item directory structure for both PDF and HTML exports, surfacing registry and entity names in the S3 path. This change applies only to new backups.
Multi-Image Viewer [2026.9.DOC.4]
Adds a new Multi-Image viewer component for the Notebook. When a user attaches 2 or more images to a notebook entry, template, or document, the images group side by side in a layout they can reorder, resize, and annotate.
Desktop Sync publishes grant read access to all users [2026.9.DOC.6]
Publishing a Desktop Sync result now makes it readable by all users in the tenant. Previously the publish step could leave the artifact readable only to a subset of users, blocking collaborators from opening the synced output.
Enable Entry Sections for all users [2026.9.DOC.7]
Replaces the old Entry Days with Entry Sections
Added a search filter for Fill-In Only Entry Templates [2026.9.DOC.8]
Added a new filter for global search that allows filtering Entry Templates by Fill-In Only Templates vs. Modifiable Templates.
Template publishing now includes review and approvals workflow [2026.9.DOC.9]
When template publishing is enabled, it is now possible to require that entry templates and subtemplates have to be approved before they can be published. Reviewers can approve, reject, and users with appropriate permissions can publish when ready. Authors and reviewers can also create test entries from a draft to confirm changes before submitting for review.
To enable this feature contact Benchling Support
Excerpts default to full width [2026.9.DOC.10]
New excerpts now default to full width when viewing an excerpt in an entry and in exports.
Item CSV exports apply per-item linked redaction [2026.9.DOC.11]
Item CSV exports now apply linked-item redaction on a per-row basis instead of using a single export-wide decision. Rows that reference items the exporter cannot read redact those columns individually while other rows continue to export their linked data.
Settings gear dropdown added for template collections [2026.9.DOC.12]
A settings gear dropdown is now available for template collections in the Projects & Templates sidepanel, giving users a single entry point for configuring a collection without opening its full settings page.
Template Collections in Workspace on for all [2026.9.DOC.13]
Templates can now be opened as a workspace tab, enabling them to be edited in context of other templates and supporting resources.
Biologics
Benchling Biologics gives antibody R&D teams a shared, structured foundation to design, register, and characterize proteins, connecting the wet and dry lab without switching tools or losing context.
Protein entity structure predictions [2026.9.BIO.1]
Users can now predict structures for protein entities using all of the structure prediction models offered in Benchling. When kicking off these predictions, Benchling automatically identifies the chains to use and the multiplicity of each chain based on the protein's format. This feature is available on the 3D structure tab on protein entities and can also be launched via Chat for one or many proteins at once. (A protein entity is not yet an available input from the Model Hub page directly.)
Protein validation accepts identical CH3 domains in the same chain [2026.9.BIO.2]
Protein registration and protein entity validation previously validated that any/all CH3s in asymmetric antibodies were unique. Now, these validation checks only ensure that interacting CH3s (e.g. CH3s on either side of an asymmetric Fc) are distinct to proxy for including the appropriate charge pair or knob/hole mutations. Other CH3s (e.g. CH3s within the same chain) can be identical.
scFc center in custom formats [2026.9.BIO.3]
Added a new center, scFc, available to be used in custom protein formats. This center includes standard Fc domains joined together in one chain as H-CH2-CH3-L-CH2-CH3.
Antibody format detection available for all Biologics customers [2026.9.BIO.4]
Benchling has added a skill to Benchling AI that determines the antibody format of a protein from its chain sequences. To use this skill, ask an agent to detect the formats for your proteins and give it the chain sequences that make up each one. We scan each protein and determine which format it is, then return the detected format with a suggested VERITAS name. The agent also restructures each protein for data import – one spreadsheet per format – with the chains arranged in the right columns with the right headers to match the detected format.
Consistent protein glyph generation in AI [2026.9.BIO.5]
Benchling has added a feature to Benchling AI that renders the glyph for a protein’s format in Chat and in Canvases whenever Benchling AI describes a protein visually. This keeps the visualization consistent with what users are seeing throughout Benchling, whether on an protein entity in registry or in an AI tool.
Biologics schemas (e.g. chains, domains) can be assigned to existing entities [2026.9.BIO.6]
Users can now assign Biologics component schemas, like domain and chain schemas, to existing entities. Previously, these schemas could only be set when creating a new entity, which blocked workflows like importing a sequence and then adding the schema afterwards.
This feature is not on by default, it is enabled by admins.
Chemistry
Plan reactions, capture structured chemistry data, and keep every molecule linked to the experiments and results that follow.
Search reactions by molecule name or lot name [2026.9.CHEM.1]
Users can now search reactions by molecule name or molecule lot name using the search bar. Reactions matching the molecule or lot query appear in the reaction search results.
Salt and Solvent libraries merged into a single Salt & Solvate library [2026.9.CHEM.2]
The chemistry Salt and Solvent libraries have been merged into a single "Salt & Solvate" library, and the Solvent library label has been renamed to Solvate. Existing entries continue to be accessible from the combined library.
Register new molecules from the reaction table [2026.9.CHEM.3]
Users can now create and register molecules directly from the reaction table in one workflow. Registration is optional and is controlled by a checkbox at the bottom of the molecule creation window.
Edit molecules linked to a submitted reaction [2026.9.CHEM.4]
Users can now edit the structure of a molecule that is linked to a submitted reaction. Reaction diagrams redraw automatically when a table re-sync picks up a new molecule version, so users see the current structure in the reaction diagram.
In Vivo
Manage in vivo studies with modern software designed for data capture, collaboration, and insights.
Measurement cells lock on deceased animals [2026.9.IV.1]
In Vivo now locks measurement cells on data tables once an animal is marked deceased. Collection and edits against animals that are no longer in the study are blocked at the cell.
In Vivo animal import template simplified [2026.9.IV.2]
The In Vivo animal import template no longer includes the redundant animal number column, and the animal name column header is now labeled Name for faster mapping during import.
Dosing report includes dosed-by name [2026.9.IV.3]
The dosing report now includes the dosed-by user's name, giving reviewers immediate visibility into who performed each dose without opening the record.
Expanded audit log coverage for studies [2026.9.IV.4]
The audit log now captures a broader set of study activity, giving you a more complete record for compliance and traceability. Events now logged include study creation and updates, approval and status transitions, team membership changes, preset changes, execution timezone changes, and graph and group settings changes.
Animal profile now displayed metadata [2026.9.IV.5]
Users can view animal metadata from the animal profile under a new tab.
SAML self-serve setup clarifies errors and certificates [2026.9.IV.6]
In Vivo SAML self-serve setup now shows a specific error page when expected SAML claims or attributes are missing, surfaces the certificate expiry date on the SAML tab, and pre-populates Audience and Issuer fields while still allowing manual edits. First-time SAML configuration is easier to complete and troubleshoot.
[Open LA] In Vivo animal metadata validation [2026.9.IV.7]
In Vivo studies now support animal metadata validation, so tenants can require specified metadata fields on animals in a study.
To enable this feature contact Benchling Support
[Open LA] Dosing in Sampling sheets [2026.9.IV.8]
The sampling sheet is now a real dose-registration surface. Users assign treatments to groups at sampling sheet creation, technicians enter weight and get an auto-calculated dose volume, and clicking Administer records the dose. Dosing time drives calculated sample collection times so that the samples are planned according to actual recorded dose times.
To enable this feature contact Benchling Support
Registry and Inventory
The Registry and Inventory combine to form an extremely powerful solution for life sciences R&D samples. The Registry can be used to model and log any biological entities while the Inventory provides an interactive digital window into the physical storage system of your lab.
Chat creates matrix plates [2026.9.RI.1]
User can use Benchling AI chat to create matrix plates now in addition to well plates.
Chat can create inventory boxes [2026.9.RI.2]
Users can now ask Benchling AI chat to create boxes
[Open LA] Using Chat for Inventory actions [2026.9.RI.3]
Several features are available in Limited Availability that allow end-users to make Inventory changes using Benchling AI. Users can now ask Benchling AI chat to perform the following actions:
- Archive containers, boxes, plates, and locations, with each request confirmed by the user before it is applied.
- Transfer samples into containers or plate wells, optionally specifying the quantity to add.
- Move one or more containers to a new storage location.
- Create one or more containers with a chosen schema, name, barcode, location, project, and schema field values.
- Look up container schemas and their fields, so it can pick the right schema when creating or filling containers.
To enable any of these features, contact Benchling Support and specify which of the bulleted capabilities you would like turned on.
Automation
Benchling Automation connects your lab instruments to your scientific record, runs analysis automatically, and feeds every result directly into the next decision, on whatever hardware you already run.
A step within automation design that allows the user to generate plate records from the barcodes or API IDs of inventory plates. [2026.9.AUTO.1]
A new Datasets to Plates step is available in the pipeline builder. Users can now add this step to convert dataset rows into plates as part of their pipelines. These plate records can then be used for transfers or analysis. Users must provide the API ID or barcode of the inventory plate.
Create, Edit and Publish Automation Templates is now GA [2026.9.AUTO.2]
Create, Edit and Publish Automation Templates is now GA
You can now build automation templates directly in Automation Designer, using the same flowchart editor as analyses. Instead of saving a finished analysis as a template, you design the template step by step, test it, and publish it when it's ready for your team More information is available in our Help Center here.
Automation templates support descriptions [2026.9.AUTO.3]
Automation templates now accept a description. Template configurer can document the purpose of a template so that users can pick the right one when creating runs.
Analysis' Apply Template panel supports auto-mapping of tables & columns [2026.9.AUTO.4]
Applying a Template from an existing Analysis now supports auto-mapping of tables and columns of Datasets. This will be configured when a user creates an Automation template with Variable source data
Register and update entities from Automation Designer are now GA [2026.9.AUTO.5]
You can now write the results of an analysis straight back to the Registry. Two new output steps in Automation Designer let you register new entities, or update existing ones in bulk, from any dataset in your analysis. You no longer need to export a file and re-import it through bulk import.
Schemas
Schemas are the structural foundation for how your information is categorized in Benchling. They are categories created specifically for your tenant that can represent materials, data, storage methods, and tasks.
[Open LA] Container schemas allow field changes for required fields [2026.9.SCH.1]
Customers can choose to opt in to be able to mark an existing field as required, or add a new required field, on a schema that already has containers, without a backfill. Containers that lack a value for the newly required field surface a missing-required-field warnings instead of blocking the schema change.
To enable this feature contact Benchling Support
Application Core
To create a seamless Benchling experience, our platform must be consistent, compliant and predictable. The Application Core provides the central capabilities to create a smoother user experience.
Setting for What's New Tenant-wide Access in the TAC [2026.9.AC.1]
Tenant admins can now enable or disable tenant-wide access to the What's New feature from the Tenant Admin Console. When disabled, the What's New button on the homepage is only visible to tenant admins.
This feature is not on by default, it is enabled by admins.
EXP ID available in Homepage My Work entries [2026.9.AC.2]
The EXP ID is available in the Homepage My Work section for entries.
Updates to visual styles of modals, buttons, and inputs [2026.9.AC.3]
This update modifies the style of most Benchling modals, buttons, and inputs to use more rounded corners and softer shadows, along with other minor visual cleanup. This brings them in-line with our latest AI features, to improve consistency and visual clarity. This change is visual styling only, no functional or feature changes.
Developer Platform
Our developer platform is fundamental to centralizing and standardizing all your R&D data. Benchling's APIs and Events are built to match the flexibility and speed of modern life science R&D.
Benchling Apps tenant limit raised [2026.9.DP.1]
Previously, tenants could only have up to 100 Benchling Apps in their Developer Console. With this change, this limit has been raised to 500 to allow creation of more Benchling Apps.
New developer docs at docs.benchling.com [2026.9.DP.2]
All Benchling developer platform documentation is now available on docs.benchling.com. This includes references and guides for APIs, SDKs, webhooks, Warehouse, and In Vivo. The new docs also include an AI assistant and interactive SDK examples.
Review process V3 API promoted to stable [2026.9.DP.3]
ReviewProcess, ReviewProcessVersion, and ReviewProcessStage are now stable in the V3 API. The GET /review-process, /review-process/{id}, /review-process-stage, and /review-process-stage/{id} endpoints no longer require the EARLY-ACCESS header, and ReviewProcess can be referenced from other stable types.
Template collection V3 API promoted to stable [2026.9.DP.4]
The GET /template-collection/items and /template-collection/{id} endpoints is now stable and collections can be filtered by name, ID, owner, archive status, and created or modified time.
In-app notifications for completed chats [2026.9.DP.5]
A new notification bell now displays unread notifications about completed chats and user @-mentions in documents. Notifications are default "on" and preference settings are coming soon.
Test order plate well associations validate well position in API [2026.9.DP.6]
Adding a plate well association to a test order now verifies that the submitted well_position is a valid coordinate for the plate's dimensions, in addition to the existing check that the plate is a fix plate. Invalid positions are rejected at the API with a clear error.
Test order result updates validate linked assay results [2026.9.DP.7]
Updating a test order result through the V3 API now accepts only readable assay results whose schema matches the result schema defined on the linked test definition. This validation previously only ran via worksheet-based updates.
Test order creation validates test definition and dilution factor [2026.9.DP.8]
Creating a test order through the API now validates that the referenced procedure has a published flowchart version, that the test definition exists in that published version, and that the dilution factor is allowed by the published test definition. The validation closes gaps with the frontend's existing enforcement in preparation for promoting the test order V3 input type to stable.
Test order V3 endpoints and inputs are stable [2026.9.DP.9]
Test order endpoints and their request inputs are now promoted to stable on the V3 REST API. Integrators can build against test order create, update, and listing endpoints without the beta designation.
Stable OligoConjugate and OligoDuplex v3 endpoints [2026.9.DP.10]
OligoConjugate and OligoDuplex endpoints were promoted to stable in the v3 API.
Stable v3 DNASequence, RNASequence, AASequence, RNAOligo, DNAOligo, and NucleotideAlignment endpoints [2026.9.DP.11]
DNASequence, RNASequence, AASequence, RNAOligo, DNAOligo, and NucleotideAlignment endpoints were promoted to stable in the v3 API.
Stable Protein and ProteinFormat v3 endpoints [2026.9.DP.12]
Protein and ProteinFormat endpoints were promoted to stable in the v3 API.
In Vivo API endpoints promoted to general availability [2026.9.DP.13]
Several endpoints are now generally available (GA) alongside their existing beta versions: Study Measurements, Study Code, Study Animal, Study Sample Plan, Study Planned Animal, Study Dosages, Study Observations, Study Samples, Observation Glossary, and Sample Glossary. Webhooks have also been added for Team Glossary Observation and Team Glossary Sample. The Study Task Specs beta endpoint has been removed because its GA version has been available for some time, so please switch to the GA endpoint if you still use the beta.
V3 API supports bulk container concentration updates [2026.9.DP.14]
The V3 API now supports updating content concentration on multiple containers in a single request through a stable bulk endpoint.
Promote V3 API endpoints for File, ScratchFile, Dataset, and DataFrame to stable [2026.9.DP.15]
The V3 API endpoints for Files, Scratch Files, DataFrames, and Datasets are now generally available.
CustomEntitySchema V3 API promoted to stable [2026.9.DP.16]
CustomEntitySchema is now stable in the Benchling V3 API, allowing reads and writes of this type without the beta stability caveat.
V3 API schema float field renamed to decimal [2026.9.DP.17]
When creating or updating schema fields in the V3 API, the field type FLOAT is now DECIMAL. This aligns with the naming of this field type throughout the rest of the product.
V3 API blob link field renamed to attachment link [2026.9.DP.18]
When creating or updating schema fields in the V3 API, the field type BLOB_LINK is now ATTACHMENT_LINK. This aligns with the naming of this field link type throughout the rest of the product.
V3 API assay result link field renamed to result link [2026.9.DP.19]
When creating or updating schema fields in the V3 API, the field type ASSAY_RESULT_LINK is now RESULT_LINK. This aligns with the naming of this field link type throughout the rest of the product.
V3 API assay run link field renamed to run link [2026.9.DP.20]
When creating or updating schema fields in the V3 API, the field type ASSAY_RUN_LINK is now RUN_LINK. This aligns with the naming of this field link type throughout the rest of the product.
V3 API assay request link field renamed to request link [2026.9.DP.21]
When creating or updating schema fields in the V3 API, the field type ASSAY_REQUEST_LINK is now REQUEST_LINK. This aligns with the naming of this field link type throughout the rest of the product.
Fieldset V3 API promoted to stable [2026.9.DP.22]
Fieldset is now stable in the Benchling V3 API, allowing reads and writes of this type without the beta stability caveat.
Schema field definitions V3 API promoted to stable [2026.9.DP.23]
All schema field definitions are now stable in the V3 REST API, allowing reads and writes of these under all stable schema types.
Unit and UnitType V3 APIs promoted to stable [2026.9.DP.24]
Unit and UnitType are now stable in the Benchling V3 API, allowing reads and writes of these types without the beta stability caveat.
Result schema V3 API supports setting legal text dropdowns on decimal fields [2026.9.DP.25]
The results schema V3 API now supports reading and writing legal text dropdowns on decimal fields. Legal text dropdowns allow users to enter values from a dropdown in addition to numbers.
Schema V3 API supports simplified field definition updates [2026.9.DP.26]
The schema V3 API set-field-definitions endpoint now supports partial updates. Only the field's id and type discriminator is required; any omitted properties are left unchanged, so callers no longer need to resend every required field value when editing one property.
V3 API supports policies and default schema configurations [2026.9.DP.27]
Policies and default schema configurations are now reachable through the V3 API. Integrations can manage these resources programmatically alongside the rest of the permissions and schema surface.
Bioprocess
Our Bioprocess offering helps with process development studies, including non-In Vivo workflows. It includes tools for tracking, managing, and analyzing experiments like process development studies, recipes, and studies.
New recipe and assay run designer [2026.9.BP.1]
This release provides a new recipe and assay run designer that visualizes how condition replicates connect across unit operations.
Custom method available for creating retest test orders [2026.9.BP.2]
A custom method now exists for creating retest test orders. Integrators can create retest orders through this method instead of hand-rolling the required field values.
Table Mapping with Condition-Replicates [2026.9.BP.3]
When using Table Mapping in Worksheets that contain a Condition-Replicate column, the values of the Condition-Replicate column will also be mapped to the destination.
PipeBio
Benchling's cloud-based bioinformatics software for antibody/peptide screening & drug development. Conduct BCR, TCR, VHH, scFv, or peptide sequence analysis and integrate your workflows with Benchling's core platform.
Split document by unique column values added [2026.9.PB.1]
Users can now split a sequence document into one document per unique value of a selected column, such as a barcode or sample column.
Allowed barcode list added to Trim UMIs/barcodes [2026.9.PB.2]
Trim UMIs/barcodes from reads now supports a list of allowed UMI or barcode values and discards reads whose value is not listed, optionally saving them to a separate document.
Jobs table filtering by parameter values added [2026.9.PB.3]
Users can now filter the Jobs table by job parameter values to find every job that ran with a specific setting.
Settable seed for ProteinMPNN and Variant Generation added [2026.9.PB.4]
Users can now set a random seed on ProteinMPNN and Variant Generation jobs to reproduce a previous result by re-running with the same seed.
Job dialogs now group output settings under Advanced settings [2026.9.PB.5]
Job dialogs now place the output name and output folder in a collapsible Advanced settings section, keeping the main form focused on analysis options.
Identifiers now shown in a monospace font [2026.9.PB.6]
Organization, user, job and document IDs and similar identifiers in settings, admin and jobs tables now use a monospace font, making them easier to read and compare.
Output folder protection for unfinished jobs added [2026.9.PB.7]
Folders can no longer be deleted while a queued or running job writes results into them. The message lists the blocking jobs so users can cancel them or wait.
Large compressed FASTQ imports now run faster [2026.9.PB.8]
Imports of large compressed FASTQ files now read sequences while decompressing, saving several minutes on files with 15 to 30 million reads.
Add to Benchling preview headers now show schema paths [2026.9.PB.9]
The Add to Benchling dialog preview now shows each mapped column's full schema path in a single header row, making deeply nested schema mappings easier to read.
Faster storage for large jobs added [2026.9.PB.10]
Jobs that process large inputs now use higher-throughput disk storage, so disk-heavy steps such as writing and uploading results finish faster.
Schematic layout for three-sample Venn diagrams added [2026.9.PB.11]
Three-sample Venn diagrams in Equal area circles mode now use a symmetric schematic layout when every overlap region is present, so small intersections stay visible.
Faster switching between document tabs [2026.9.PB.12]
Switching back to an open document tab now shows it immediately from memory and refreshes it in the background only if the document changed.
Dialog field errors now shown under each field [2026.9.PB.13]
Validation errors in dialogs now appear directly under the field that needs attention, using one consistent style across dialogs.
Linclust clustering engine for large inputs added [2026.9.PB.14]
The Cluster and Cluster and compare multiple samples dialogs now offer Linclust as a clustering engine, which clusters very large sequence documents faster than the default engine.
Folder tree drop zone for file uploads added [2026.9.PB.15]
Dragging files over the folder tree now shows a drop zone, so users can see where the files will be uploaded before dropping them.
PipeBio aligner for germline annotation added as beta [2026.9.PB.16]
The Annotate dialog now offers the PipeBio aligner, a faster germline search whose results closely match BLAST, as a beta search algorithm with an explanatory note and documentation link.
Bug Fixes
Benchling AI
- Fixed an issue where credit limits were mistakenly being enforced in non-Production GxP tenants
- Fixed an issue where enabled AI connectors failed after granting chat approval.
- Prediction batches now appear as failed after timing out
Documents
- Paste as text preserves destination formatting
- Structured table columns preserve units
- Prevent rounding of scientific notation values in Structured tables
- Apply None color to selected object in Image Annotation
- Rejected template reviews end read-only reviewer draft access
- Excel to PDF exports honor page setup
- Draft-version templates no longer appear in Completed on the Homepage
- Partial export email item IDs preserved on repeats
- Resolve text to Entity link after creation
- Set precision on columns with units
- Keep first keystroke when typing with IME
- Allow collapsing all tables in read-only entries
- Bulk exports of AV-blocked files keep their sentinel filenames
- Folder exports skip nested selected folders
- Capability checks enforced on print and export editor
- Exports omit linked item names the user cannot read
- Clearer document-export API errors and audit records
- Fixed missing characters in audit log and location history PDF exports
- Export audit log works for template collections in search
Molecular Biology
- Genome dropdown recovers when Ensembl is unavailable
- HELM computed field keeps reverse-strand modifications
- Fixed interacting Autofill and Done buttons on the metadata tab of a sequence entity
- Amino acid validation properly rejects invalid letters
- Improved performance when loading protein formats
Biologics
- Legacy antibody annotations available on non-Biologics AA sequences
In Vivo
- Fixed incorrect days used for measurement columns when creating datatables from template
Registry and Inventory
- Audit log records final registry ID after API registration
- Inventory schema field type changes no longer time out
Workflows
- Workflow task-definition picker loads large libraries
- Task table edits persist when output tables are submitted
- Request table loads with deleted sample links
- Snapshot fields on workflows support tooltips
Automation
- Transient Run ingest analysis failures resolved
- Published automation templates duplicate cleanly
- Calculate interpolation step validates model variables input
- Send to Document step shows informative error for missing dataset
Schemas
- Archived schema fields cannot be modified in place
Access Management
- Review prompt restriction
- Admin user dashboard loads for users with open reviews
Application Core
- Paginated audit log scales to large histories
- Procedure parameter config export excludes parent values
- Request attachment removal shows readable audit label
- Save button re-enabled after cancelling move reason modal
- Audit log hides barcodes of unreadable affected objects
Developer Platform
- Draft review changes hidden from template read-only viewers
- Review process listings exclude archived items by default
- Review history reports completed cycle actions
- v2 DNASequence PATCH endpoint treated null fields as absent
- StudySchema create returns 400 when studies are disabled
- Hide inaccessible review processes via v3 APIs
- Fixed several issues where creating an Entry with an Entry schema would error.
- V3 document-export endpoints hardened for polling and input validation
- V3 document export returns 404 for missing or unreadable items
- V3 document export status returns 404 for unknown task IDs
- Apply fieldset V3 API returns Not Found for missing schemas
- DropdownLinkFieldDefinition listing endpoint removed from V3 API
Bioprocess
- Applying an analysis template respects step limit
- Procedure run plans cannot be created outside execution phase
- Fixed Procedure Run Plan writes
- Fixed an issue when deleting a study plan and adding a study plan
- Study schema editor matches backend study-type enablement
- Study Visualization overview toggle closes panel on reclick
- Improved performance loading and saving Worksheets
- Improved table performance for Worksheets with Cross-Step Data References
- Improved performance of loading steps in Worksheets
- Procedure run plan pages render reliably
Development
- Recipe runs render without template collection write access
- Equipment schema model cell handles pasted any-entity columns
- Material input default values show validation errors after schema switch
- Recipes method editor save and blank row behavior
- Recipe config tables keep equipment in saved order
- Mixture Prep Instructions with referenced Recipe
PipeBio
- Fixed bug merging reads with duplicate names
- Fixed bug exporting alignment views with trees
- Fixed bug sorting text columns in sequence documents
- Fixed bug showing folder tree column separators
- Fixed bug trimming 3' UMIs and barcodes
- Fixed bug hiding alignment tree when columns are sorted
- Fixed bug displaying both chains in Extract sequences
- Fixed bug merging assay data with padded keys
- Fixed bug finding hidden columns in column search
- Fixed bug explaining missing scaffold chains in Correct Liabilities
- Fixed bug running large ProteinMPNN variant requests
- Fixed bug submitting Subsample sequences jobs
- Fixed bug explaining chain requests on unpaired documents
- Fixed bug completing jobs on long-named documents
- Fixed bug validating required Add to Benchling fields
- Fixed bug showing gaps in chromatogram traces
- Fixed bug sorting unsaved rows in sequence tables
- Fixed bug adding empty germline history versions
- Fixed bug opening node options in cluster network charts
- Fixed bug with job input downloads after network interruptions
- Fixed bug restoring Re-array plates in Assay data
- Fixed bug matching Add to Benchling values to entities
- Fixed bug with the chain selector on unpaired documents
- Fixed bug showing the template dropdown in Create workflow
- Fixed bug loading comparison charts without an x-axis
- Fixed bug validating unsupported workflow Splitter layouts
- Fixed bug closing and scrolling multi-select dropdowns
- Fixed bug with the join icon in Merge assay data
- Fixed bug selecting germlines in the Annotate dialog
- Fixed bug clustering on regions past the sequence end
- Fixed bug running diversity selection on identical sequences
- Fixed bug redrawing the sequence viewer after zooming out
- Fixed bug exporting horizontally scrolled tables to PNG
- Fixed bug adding annotations to unannotated sequences
- Fixed bug restoring the personal settings link
- Fixed bug aggregating values on the id column
- Fixed bug uploading many unsupported files