Daily Digest

Daily Digest – 24 June 2026

2026 Capabilities Progress

  1. Page generation — Canvas AI work continued on exposing block properties to agents and improving the UI during page builds, including locking the canvas interface during generation [77][81][99]. New mockups for AI page creation were delivered to define the user experience [43].
  2. Context management — CCC merged performance optimizations for conditional subcontext and usage tracking, closed issues for document loader support for PDFs, markdown, and web pages, and advanced toward beta3 release [108][110][113][114][115][122][124][125]. Discussion began on extensible role architecture for context items [118].
  3. Background agents — New agent skills for automated configuration of settings, guardrails, and automators are actively in development [6][7][17][18]. Work advanced on conditional subcontext provider calls and scheduler feature improvements [108][126].
  4. Design system integration — Canvas work continues on component instance updates and translation propagation with alternate architecture approaches under review [83][98]. Tool API planning continues for SDC component management capabilities, though no implementation progress occurred today [72].
  5. Content creation and discovery — Batch embeddings work progressed across multiple providers and AI Search to improve indexing performance [3][37][56][58][127][128]. Interactive refinement capability for field widgets advanced significantly [5][24][129][130]. Chat processor plugin work continued with open issues and merge requests in progress [15][16][38][39][40][41].
  6. Advanced governance — Canvas conflict detection was placed behind a feature flag to support controlled rollout, with multiple merge requests in progress on conflict resolution interfaces [76][87][89][90][91][92][95]. Mockups for AI review notifications and on-edit review were delivered [44][45].
  7. Intelligent website improvements — No direct progress today.
  8. Multi-channel campaigns — Mockups for AI page translations were completed, defining the user experience for multi-language content workflows [49].

Shipped

  1. AI module patch releases prepared — Three releases (1.2.17, 1.3.8, and 1.4.3) are in preparation across supported branches [13].
  2. Simplified AI provider selection — The provider configuration form element is being integrated into existing workflows to make model selection easier for site builders [2][29][30].
  3. Automatic vector database defaults — Vector database providers can now automatically become the default when none is configured, reducing manual setup steps [8][26][27][31].
  4. Long-running content support for AI Search — Content chunking for large articles and documents was backported to the 1.x branch, ensuring AI Search can handle content of any length [10][32][34].
  5. Automators moved to structured output — This architectural change improves reliability and consistency of AI-generated content [11][35].

Ongoing

  1. Batch embeddings for faster indexing — Work is progressing to process multiple embedding requests together, reducing processing time and API costs for high-volume sites [3][37][56][58][127][128].
  2. Agent configuration skills — New capabilities are under development to automate setup of AI settings, guardrails, and automators through natural language [6][7][17][18].
  3. Interactive content refinement — A feature allowing users to iteratively improve AI-generated content with follow-up instructions is under review [5][24][129][130].
  4. Chatbot plugin integration — Work continues on integrating the ChatProcessor plugin system into the AI Chatbot block [15][16][38][39][40][41].
  5. Canvas translation propagation — The team is evaluating multiple approaches to ensure component updates flow correctly to all language versions [83][98].

AI (Artificial Intelligence)

The team delivered several important updates in the past 24 hours, closing multiple feature gaps and advancing the project toward its next stable releases. Three patch releases (1.2.17, 1.3.8, and 1.4.3) are being prepared across supported branches.

Key improvements now complete include streamlined configuration interfaces for model selection, better defaults for vector database providers, and support for handling large content in the AI Search feature. The provider configuration form element was integrated into existing workflows, making it easier for site builders to select AI models. Vector database providers can now automatically become the default when none is configured, reducing manual setup. The team also backported long-running content chunking to the 1.x branch, ensuring AI Search can handle articles and documents of any length. A significant structural change, moving Automators to structured output, was merged, improving reliability and consistency of AI-generated content.

Work is actively progressing on several features for upcoming releases. Contributors Artem Dmitriiev (a.dmitriiev) and Ann Mary Sruthy (annmarysruthy) are building agent skills that allow automated configuration of settings, guardrails, and automators. David Galeano (gxleano) and Ahmad Khader (Ahmad-Khader) are integrating chatbot plugin architecture. A team led by Ahmad Khalil (ahmad-khalil-imagex) is working on batch embeddings to accelerate indexing performance. Documentation improvements and a potential regression in the CKEditor integration are also under review.

How can I help on this project?

  • Approve the three upcoming patch releases (1.2.17, 1.3.8, 1.4.3) once testing is complete to deliver recent improvements to users.
  • Prioritize the CKEditor regression investigation to confirm whether earlier functionality has been lost.
  • Support documentation work by reviewing and approving the project page update to help users understand release options.

View project on GitLab: ai

Drupal AI Initiative

The Drupal AI Initiative made significant design progress over the past 24 hours, completing a suite of visual mockups that define how content editors and site administrators will interact with AI-powered features. Niels Aers (nielsaers) delivered eight interactive mockups covering key editorial workflows: AI page creation, review notifications, on-edit content review, AI-powered search, field automation, search and replace, translation, and dashboard setup. These mockups establish the user-experience foundation for upcoming features and provide the team with a shared visual reference.

On the development front, Marcus Johansson (Marcus_Johansson) opened work to recruit beta testers for the Module Finder, a tool that helps developers discover existing community modules before building custom code. The team also began exploring skills for hosting providers, starting with DevPanel API research to enable automated environment setup and deployment through natural-language instructions.

How can I help on this project?

  • Support the Module Finder beta recruitment effort by connecting the team with potential test users from your network.
  • Prioritize feedback cycles for the newly delivered mockups to accelerate design-to-development handoff.
  • Evaluate partnership opportunities with hosting providers to ensure API access for deployment automation skills.

View project on GitLab: ai_initiative

AI Integration - ECA

The AI Integration - ECA module saw one new issue opened during the reporting period. Work has been assigned to examine how users currently configure structured output settings when using chat functionality. The current approach requires manually entering JSON code, which can be error-prone and difficult to review once saved. The assigned task is to explore whether a more user-friendly configuration interface is feasible, one that would validate inputs automatically and make settings easier to understand and maintain.

This work is labeled as part of the "Road to 1.0.0" effort, indicating it contributes to preparing the module for stable release. The issue has been assigned to Jürgen Haas (jurgenhaas) for investigation.

No code changes were merged or committed during this period. This level of activity is typical for a module in active development approaching its first stable release.

How can I help on this project?

  • Ensure the developer assigned to this work has sufficient time allocated to investigate and implement the improved configuration interface before the 1.0.0 release deadline.
  • Consider whether user testing or feedback is needed to validate that the new interface meets administrator needs.

View project on GitLab: ai_integration_eca

OpenAI Provider

The OpenAI Provider module made strategic progress this period by keeping pace with platform changes and beginning work on efficiency improvements.

The team merged a fix to remove Dall-E 3 as the default image generation model, reflecting its deprecation by OpenAI. This update ensures users aren't defaulted to an outdated option and keeps the module aligned with current platform capabilities. Contributor Abhisek Mazumdar (abhisekmazumdar) delivered this change.

Separately, Paul Sheldrake (paulsheldrake) opened work on add support for batch embeddings, which will enable the module to process multiple embedding requests together rather than one at a time. This capability is important for sites handling larger volumes of content, as it can reduce processing time and API costs.

How can I help on this project?

  • Review and approve the pending batch embeddings work to unlock performance improvements for high-volume use cases.
  • Ensure the team has access to current OpenAI platform documentation and deprecation notices to maintain alignment with upstream changes.

View project on GitLab: ai_provider_openai

Gemini Provider

The Gemini Provider project saw development activity this period focused on expanding its capabilities. A merge request for Support batch embeddings was opened by Artem Dmitriiev (a.dmitriiev), representing nearly 400 lines of changes. This work would enable the module to process multiple pieces of content at once rather than one at a time, improving efficiency for organizations working with large volumes of content that need to be analyzed or processed through the AI service.

The merge request remains under review and has not yet been integrated into the project. No other issues or commits were recorded during this period.

How can I help on this project?

  • Arrange for technical review resources to evaluate the pending batch embeddings work, accelerating its path to completion.
  • Clarify organizational priorities around AI content processing features to help maintainers focus development efforts effectively.

View project on GitLab: gemini_provider

Microsoft Azure AI

The Microsoft Azure AI module saw continued development over the past 24 hours, with improvements made to the text-to-image functionality. This enhancement, delivered by Marcus Johansson, expands the module's capability to convert written descriptions into visual content using Azure's artificial intelligence services. For organizations using this module, this means better quality or more reliable image generation from text prompts, which can support content creation workflows and marketing automation efforts.

No new issues or merge requests were recorded during this period, indicating a quiet day for community collaboration and issue resolution. The project continues its development trajectory with incremental improvements to core features.

How can I help on this project?

  • Consider allocating resources for user testing of the improved text-to-image feature to validate its business value and identify any gaps before wider adoption.
  • Evaluate whether additional development capacity is needed to accelerate progress toward a stable release if this module is strategically important to your organization.

View project on GitLab: ai_provider_azure

AI Agents

Over the past day, the AI Agents project saw focused activity on user experience improvements and bug fixes affecting how agents interact with Drupal Views and how errors are handled in the interface.

Three merge requests were opened to address known issues. Work is progressing on reorganizing the administration menu to align with the AI module's 1.4.x structure, placing AI Agents under a Tools & Automation category for clearer navigation. A solution is in progress for preventing exceptions in tool execution from breaking the user interface, which will ensure users receive clear feedback when something goes wrong rather than encountering silent failures. Additionally, work continues on adding configurable limits on tool usage to prevent agents from entering infinite loops.

Several new issues were identified around the Views agent's ability to update display configurations correctly, and contributors are investigating schema-generation logic when agents work with complex data structures. Ahmad Khader opened an issue to improve the JSON schema editor interface, making it easier for site builders to configure structured output.

How can I help on this project?

  • Prioritize testing and code review resources for the three open merge requests to accelerate their path to production.
  • Consider allocating developer time to the Views-related bugs, which affect a core integration point for end users building content listings.

View project on GitLab: ai_agents

Tool API

The Tool API project delivered a fix that improves how the system handles and reports errors. Previously, when a tool operation failed, the system would crash while trying to read the results, hiding the actual failure message from users. This meant that when something went wrong, users saw a confusing technical error instead of a clear explanation of what happened.

Matt Glaman (mglaman) identified the problem and merged a solution within the reporting period. The fix ensures that when a tool fails, users now see the intended failure message rather than a technical exception. This brings the command-line interface in line with other parts of the system that already handled errors correctly.

Separately, there is ongoing design work on tools to save, update, and delete SDC components, though no implementation activity occurred during this period. This feature would enable programmatic management of design components once implemented.

How can I help on this project?

  • Clarify the priority and timeline for the SDC component management feature, which currently has no assigned developer.
  • Connect the team with product owners who can validate whether the planned component tools meet user needs before development begins.

View project on GitLab: tool

Drupal Canvas

The team made steady progress on translation features and infrastructure improvements over the past 24 hours.

A fix to template translation deletion was completed and merged by Kunal Sachdev, ensuring the correct route is used when removing translated content templates. This resolves an issue affecting multilingual content management workflows.

Christian López Espínola (penyaskito) closed the translation propagation merge request after significant work on ensuring component updates flow correctly to all language versions. An alternate approach is under active development that may offer better maintainability. Related work includes addressing a test reliability issue where database backends handle JSON key ordering differently, and splitting the test suite across four parallel jobs to speed up the development cycle.

To prevent translation work being disrupted, Feliksas Mazeikis (f.mazeikis) is working to place conflict detection behind a feature flag, allowing teams to control when validation rules are enforced. Multiple merge requests remain in progress on conflict resolution interfaces and backend APIs.

The Canvas AI team continued refining the page generation feature, with discussions on locking the interface during builds and improvements to how block properties are exposed.

How can I help on this project?

  • Review and approve the translation propagation architecture decision. Multiple approaches are under consideration, and executive alignment on the preferred direction would accelerate delivery.
  • Confirm the priority of conflict resolution features versus other roadmap items to help the team allocate resources effectively.

View project on GitLab: canvas

Context Control Center (CCC)

The Context Control Center (also known as AI Context) project completed planning for its upcoming beta3 release and wrapped up Sprint 12 work, while beginning Sprint 13 with a focus on release preparation and user experience improvements.

Several performance optimizations were merged this period. The team addressed database efficiency in the usage tracking system, which records how AI agents use context in production. Work also progressed on making conditional subcontext (smart context that is dynamically selected by AI based on conversation needs) more efficient and easier to manage when administrators need to turn it off.

The project expanded its document loading capabilities, completing work on a new submodule that lets users provide context from PDFs, text files, and web pages. This closed three related features: webpage context support, Markdown file support, and PDF support. The team also closed planning work for the document loader feature overall.

User-facing improvements are in progress, including better guidance for beta testers through updated project documentation and clearer warnings when editing context items that depend on unpublished parent items.

How can I help on this project?

  • Identify and connect potential beta testers from your network who could provide feedback on the expanded document loading features and overall context management workflows.
  • Review the open discussion about integrating the new Annotations module to determine if alignment or partnership would accelerate both projects.

View project on GitLab: ai_context

Field Widget Actions

The Field Widget Actions module saw significant progress on a new capability that will let users polish AI-generated content without starting over. Edward Zwart (ezeedub) opened a merge request that introduces an optional refinement feature: after the module generates content, users can open a dialog to iteratively improve it with follow-up instructions like "make it shorter" or "add more detail," then insert the final version into their content.

This addresses a key gap in the current workflow. Today, once content is generated, users must either accept it as-is or manually edit it. The new feature bridges automated generation and user control, making the module more practical for real-world editorial needs.

The work was built using standard form components to keep the architecture simple and maintainable. Refinement is off by default, so existing users won't see any change in behavior unless they choose to enable it.

How can I help on this project?

  • Prioritize review and testing of the refinement feature to accelerate its path to release.
  • Identify pilot users or content teams willing to trial the feature and provide feedback on workflow fit.

View project on GitLab: field_widget_actions

AI translate

The AI translate module experienced a newly reported issue during this period. A bug was identified where the translation system is adding unwanted HTML markup (specifically line break tags) after certain existing HTML tags in content. This issue was discovered when translating basic page content, where the AI provider is inserting extra formatting that wasn't present in the original text. The problem affects content quality and could require manual cleanup after automated translation, reducing the efficiency gains the module is meant to provide.

The issue was reported by Tom Wauman (tom.wauman) and has been accepted as a confirmed bug with normal priority. No merge requests or fixes have been submitted yet, and the issue remains unassigned. This is typical for an evolving module where issues are being identified and triaged.

How can I help on this project?

  • Allocate developer resources to diagnose and fix the HTML markup issue, as it currently has no one assigned to resolve it.
  • Prioritize testing across different content types to identify whether similar formatting problems exist elsewhere in the translation workflow.

View project on GitLab: ai_translate

Pinecone VDB Provider

The Pinecone VDB Provider module enables integration with Pinecone's vector database service, allowing the project to store and search large amounts of data efficiently. Over the past 24 hours, development has focused on improving the module's performance when handling large-scale data operations.

Rob Loach (robloach) opened a new feature request for Support batching on indexing operations, which has been assigned to him and is under review. This feature would allow the system to process data in groups rather than one item at a time, improving efficiency when adding large volumes of content to the database. This capability is already being implemented in similar components, and extending it to Pinecone will ensure consistent performance across the platform.

No code has been merged yet, as the work is in the planning and review phase. This is normal progress for feature development.

How can I help on this project?

  • Clarify priority and timeline expectations for the batching feature to help the developer balance this work against other commitments.
  • Ensure testing resources or environments are available when the implementation is ready for validation.

View project on GitLab: ai_vdb_provider_pinecone

References

[2] Use ai_provider_configuration form element in RestrictToTopic configuration form (ai, Issue)

[3] Batch embeddings for improved indexing performance (ai, Issue)

[5] Add interactive refinement capability to Field Widget Actions (ai, Issue)

[6] Add a setup-guardrails agent skill to configure guardrails, guardrail sets, and global guardrails (ai, Issue)

[7] Add a configure-ai-settings agent skill for general AI settings (default models, providers, timeout, trusted domains) (ai, Issue)

[8] Allow vector database providers to opt in as the default VDB when none is set (ai, Issue)

[10] Support long running chunking in 1.x branch for AI Search 1.x readiness (ai, Issue)

[11] Move Automators to Structured Output (ai, Issue)

[13] Release 1.2.17, 1.3.8, 1.4.3 (ai, Issue)

[15] Add AI Assistant API ChatProcessor plugin (ai, Issue)

[16] Integrate ChatProcessor plugin system into AI Chatbot DeepChat block (ai, Issue)

[17] [2.x] Add a setup-automators agent skill to configure automators on fields (ai, Issue)

[18] [1.x] Add a setup-automators agent skill to configure automators on fields (ai, Issue)

[24] Issue #3586468: "Add interactive refinement capability to Field Widget Actions Automation" (ai, MR)

[26] Issue #3586500: Get default Vdb provider if none yet selected (ai, MR)

[27] Issue #3586500: Get default Vdb provider if none yet selected (ai, MR)

[29] Issue #3586472: Use ai provider configuration form element and use proper... (ai, MR)

[30] Issue #3586472: Use ai provider configuration form element and use proper... (ai, MR)

[31] Issue #3586500: Get default Vdb provider if none yet selected (ai, MR)

[32] Non-BC support for long running chunks (ai, MR)

[34] Non-BC support for long running chunks (ai, MR)

[35] feat: #3577170 Move Automators to Structured Output (ai, MR)

[37] Batch embeddings for improved indexing performance (ai, MR)

[38] #3585077: Integrate chatprocessor plugin (1.x backport) (ai, MR)

[39] #3585075: Add ai assistant chat processor (1.x backport) (ai, MR)

[40] #3585075: Add ai assistant chat processor (ai, MR)

[41] #3585077: Integrate chatprocessor plugin (ai, MR)

[43] Mockup: AI page creation (ai_initiative, Issue)

[44] Mockup: AI review notifications (ai_initiative, Issue)

[45] Mockup: AI review on edit (ai_initiative, Issue)

[49] Mockup: AI page translations (ai_initiative, Issue)

[56] add support for batch embeddings (ai_provider_openai, MR)

[58] Support batch embeddings (gemini_provider, MR)

[72] [Tool] Create tools to save/update/delete SDC components (tool, Issue)

[76] Put conflict detection behind a dev mode flag (canvas, Issue)

[77] Canvas AI: UI/Functional update. Lock the canvas during a build and add a long-running AI-tasks UI element to the chat (canvas, Issue)

[81] Canvas AI: Expose Props of Blocks to the Agent (canvas, Issue)

[83] Draft: feat: #3591596 ALTERNATE Propagate component instance updates to content/config translations; introduce `StagedLanguageConfigOverride` to treat both similarly (canvas, MR)

[87] Put conflict detection behind feature flag (canvas, MR)

[89] Draft: feat(Conflict detection): #3591668 "Text-based Page entity representation via Canvas API" (canvas, MR)

[90] Draft: feat(Conflict detection): #3591601 "Conflict resolution via API call and layout endpoint support for published entities" (canvas, MR)

[91] feat(conflict): Add conflict resolution UI with viewport selection and route integration (canvas, MR)

[92] feat(Conflict detection): #3587526 "Detect conflicts during publishing of changes to Page entity" (canvas, MR)

[95] feat(Conflict detection): #3587526 "Client side prevention of publishing" (canvas, MR)

[98] feat(Component sources): #3591596 Symmetric translation: component instance version updates must propagate to all translations" (canvas, MR)

[99] Issues/3569120: Canvas AI: Expose Props of Blocks to the Agent. (canvas, MR)

[108] Conditional Subcontext Provider Calls (ai_context, Issue)

[110] CCC Usage Tracking Query/Save Pattern (ai_context, Issue)

[113] Add webpage (URL) context source plugin to CCC (ai_context, Issue)

[114] Add MD context source plugin to CCC (ai_context, Issue)

[115] Add PDF context source plugin to CCC (ai_context, Issue)

[118] [Discuss] Extensible "role" architecture for context items (examples, rubrics, and future types) (ai_context, Issue)

[122] Issue #3586207: Ignore conditional context when disabled. (ai_context, MR)

[124] Issue #3586207: Add conditional subcontext toggle, cap, and per-request cache. (ai_context, MR)

[125] Issue #3586209: Batch N+1 queries in AiContextUsageTracker (ai_context, MR)

[126] Issue #3577429: Make Scheduler an optional progressive feature. (ai_context, MR)

[127] Batch embeddings for improved indexing performance (ai_search, Issue)

[128] Resolve "Batch embeddings for improved indexing performance" (ai_search, MR)

[129] Add interactive refinement capability to Field Widget Actions (field_widget_actions, Issue)

[130] Issue #3547306: Add interactive refinement capability to Field Widget Actions (field_widget_actions, MR)

Period: 24 June 2026 to 25 June 2026  |  Generated: 25 June 2026 05:00 GMT

To view the full AI prompts used to generate this digest check this file: 24h_2026-06-24_prompts.txt

Looking for older digests? Browse all digests here.