Feature Request

Currently, Eddy AI offers only two controls for customizing how it responds to customer queries: Tone Response length: Even this is subjective as we cannot provide a custom value. While these are useful, they are not sufficient for organizations that want more control over the quality, format, and behavior of AI-generated responses. What we are requesting: The ability to provide custom instructions to Eddy AI, similar to how system prompts or instruction files work in other AI tools such as ChatGPT or Claude, that Eddy AI can follow before generating a response. These instructions could be used to: Enforce a specific response format or structure Define rules around what Eddy AI should or should not include in a response Require Eddy AI to perform certain checks before responding (e.g. verify the answer exists in the documentation before returning it) Handle specific queries in a predefined way rather than attempting to generate an answer by itself How it could work: A dedicated text field or configuration panel within the Customize Eddy AI settings where admin can insert custom instructions that are applied globally before every response is generated. Optionally, the ability to define query-specific instructions or standard responses for certain topics or keywords would make this even more powerful. Why this matters: Organizations using Document360 as a customer-facing knowledge base need confidence that Eddy AI is not only answering questions, but answering them in the right way, in the right format, and within the right boundaries. Custom instructions would give administrators the control needed to achieve that.

Currently, when Eddy AI credits are exhausted, users are unable to access Eddy AI Analytics and are shown the page seen in the attached screenshot. Additionally, when users purchase new AI credits, they often assume that Eddy AI will automatically be re-enabled for their knowledge base, as it was before. However, the Eddy AI toggle is turned off by default when new credits are added, requiring users to re-enable it in the settings manually. It would be helpful to display a message prompting users to enable the Eddy AI toggle from the settings page after new credits are added.

When replicating articles from one workspace to another, hyperlinks within the replicated content continue to point to the original workspace instead of updating to the corresponding articles in the destination workspace. This is a significant issue for organizations that maintain multiple workspaces for different customer bases. Readers who only have access to the destination workspace (workspace 2) cannot access the linked content because the hyperlinks redirect them to the source workspace (workspace 1), which they do not have access to — even when the linked articles have already been replicated to workspace 2. Example: Article A is replicated from workspace 1 to workspace 2 Article A contains hyperlinks pointing to Article B in workspace 1 Article B has also been replicated to workspace 2 However, the hyperlinks in the replicated Article A still point to Article B in workspace 1 Readers in workspace 2 cannot access workspace 1, so the links break for them Expected behavior: When an article is replicated to a destination workspace, any hyperlinks pointing to articles that also exist in the destination workspace should automatically update to point to the correct articles within that workspace. Impact: This is causing a poor reader experience for customers who are scoped to a specific workspace and do not have access to the source workspace.

We need a way to get Eddy AI search questions, conversations, and feedback data without manually clicking Export CSV every time. Right now there's no API for this data. The only workaround is scripting the internal export button, but that's not supported, needs a stored login, gets blocked by Cloudflare bot protection, and only gives short-lived tokens with no way to refresh them. We'd like an API endpoint to pull this data using a normal API token, so we can access it on a schedule without any of the above issues. Here's the linked FR : https://feedback.document360.com/feature-request/p/external-access-to-eddy-ai-search-queries Thanks

Currently, a Document360 project supports a single base domain, which is shared across all workspaces within that project. If a customer wants each workspace to be hosted on a completely different domain (not just a different subfolder or path on the same domain), this is not natively supported. The only current workaround is to purchase separate projects, one per required domain. We would like to request the ability to assign a distinct custom base domain to each workspace within the same project. This would allow teams managing multiple workspaces for different products, following different branding, under the same company, to maintain a single project (with shared settings, categories, and management) while still publishing each workspace under its own unique domain. Use case: Our organization has multiple workspaces under one project, each representing a different product with its own branding, under the same company. The target audience for these products may be the same in some cases and different in others. When the target audience is the same, we prefer to keep these as workspaces within one project rather than move to separate projects, so that we can view analytics for that audience in one place instead of across separate dashboards. Staying within a single project also makes it easy to replicate content across workspaces. The request is simply to allow a different base domain for each workspace, without needing to maintain separate projects (and separate subscriptions) just to achieve domain separation. Requested capability: Ability to configure a unique custom base domain per workspace within a single project. Independent SSL/DNS configuration per workspace domain. Retain shared project-level features (content replication across workspaces, analytics, user management, categories) even with different domains per workspace. This would reduce the need to purchase multiple projects solely to achieve domain-level separation between workspaces.

When a glossary term is deleted, the system currently removes both the glossary reference and the associated text from article content. This behavior makes it difficult to reuse or retain the original wording within articles. We request an option to preserve the term text by converting it into plain text instead of removing it entirely. Problem Statement In the current implementation, deleting a glossary term results in the complete removal of that term from all associated articles, including the visible text. This creates several issues: -> Loss of content within articles -> Difficulty in identifying where terms were previously used -> Manual effort required to reinsert or replace removed text Expectation When a glossary term is deleted, users can choose between: -> Remove completely -> Convert to plain text

I want to be able to report on how many articles a user contributed to that got published in x time frame. I want to see if those articles were new vs edits of existing articles. The way that it is only contributions that are shown on a user level is not enough. I also want to see the list of articles that each user contributed to. Currently when you click into the user it doesn't apply time filter that was already applied to the list view. We are trying to understand better each user's performance over a month and a quarter and it's very hard and manual to pull all these details.

Currently, glossary entries with the same Name and Term cannot be imported across different languages, as the glossary Name must be unique across the project. The customer uses an external translation service and imports glossary entries separately for each language. In some cases, the Term remains the same across languages (for example, "Widget"), while only the definition changes. The customer would like the ability to import the same glossary Name and Term across different languages and have them treated as language-specific entries. The glossary import should also provide an option to specify the target language and name of the glossary in the csv.

Currently, headings located inside Tabs cannot be linked using anchor links. It would be helpful to support linking directly to headings within Tabs, enabling users to navigate to specific sections inside tabbed content.

I want to be able to add comments to glossary entries, for example. This will help me add context to why and how an entry was created. For example, I'm just describing a UI feature and saw that there is a glossary entry for its counterpart. However, I can no longer figure out how I created this glossary entry. Having this info would help me easily create the new entry with the same surrounding context, instead of guessing and possibly causing a mismatch. This could be supplemented by metadata, like created/updated by or a status dropdown for Draft, Approved, etc.