Reference another chat for context

The ability to reference another chat for context to the current one would be very useful. ChatGPT provides this feature, so this is a common need. Forking can provide a way to do this, but then the previous chat is still cluttering up the new one. So forking is useful in some cases, but not having to scroll past a long chat to get to the new part would be useful too.

Add chats to knowledge base.

The ability to easily chats to the knowledge base (or else an AI summary of them) would be very useful. ChatGPT does this automatically - tailoring responses based on previous chats.

Support Anthropic Claude "Agent Skills" Integration via Custom Model Parameters

Description: Please add support for Anthropic's new Agent Skills (Skills API) in TypingMind’s custom model configuration. This would allow users to specify the container parameter, custom Skill IDs, and required beta headers (such as code-execution-2025-08-25 and skills-2025-10-02) when configuring Claude models—enabling full usage of Anthropic's "Skills" within the TypingMind chat interface. Key Features Needed: Ability to set container (with skills) parameter in custom model advanced settings Support for the required beta headers for code execution and Skills Option to select and manage custom skills (skill IDs) for use in Claude-powered chats UI support for Skill versioning and multi-Skill workflows (Optional) Files API integration for downloading/uploading generated files Benefits: Unlocks Advanced Automation: Use pre-built and custom Anthropic Skills for Excel, PowerPoint, PDF, scripting, and more, directly from the TypingMind interface. Empowers Agents & Workflow Building: Enable automations, code execution, custom document generation, and business process integration with Claude’s secure code sandbox. No Need for External Scripting: Use Anthropic’s Python client features from within TypingMind—no extra coding or external scripts required. Keeps TypingMind Competitive: Stay at the leading edge of LLM platform capabilities as Anthropic deploys more enterprise co-pilot and tool automation use cases. Scalable Across Teams: Organizations can administratively manage and evolve custom Skill libraries for all Claude-powered agents organization-wide. Thanks for considering this high-impact upgrade!

Rearrange pluggins!

Is there a way we can rearrange pluggins in the chat and the plugins page similarly like models? This would be a really cool and important feature for organizing. Especially when we have alot of pluggins overtime it can be confusing if we don't have the option to rearrange them as we want. Thanks!

Add a Agent Skills to Typingmind - https://agentskills.io/

Agent skills are MCPs on steriods, can you please add support for skills?

An option to disable the new "Smart Default Model Selection"

My use case: I always use multi-models by default. I always switch models (in fact, model sets) in a single chat. If a prev chat ended with “GPT 5.1“ as the last model selection, it does not in any sense mean I want to start the next chat with “GPT 5.1“. I’d like an option to go back to the static multi-default models.

High-Res (2K/4K) Output for Gemini 3 Pro Image

Please add a parameter to allow image generation at 2K and 4K resolutions for Gemini 3 Pro Image. The current 1K limit holds back the model's potential for professional high-detail work. Having native high-resolution support would make Gemini a much stronger competitor against other image models and eliminate the need for external upscaling tools.

Implement an "LLM Council" in multiple-LLM chat

Based on karpathy’s idea : https://github.com/karpathy/llm-council The idea is that instead of asking a question to your favorite LLM provider (e.g. OpenAI GPT 5.1, Google Gemini 3.0 Pro, Anthropic Claude Sonnet 4.5, xAI Grok 4, eg.c), you can group them into your "LLM Council". A query is sent to multiple LLMs, it then asks them to review and rank each other's work, and finally a Chairman LLM produces the final response. In a bit more detail, here is what would happen when you submit a query: Stage 1: First opinions. The user query is given to all LLMs individually, and the responses are collected. The individual responses are shown in a "tab view", so that the user can inspect them all one by one. Stage 2: Review. Each individual LLM is given the responses of the other LLMs. Under the hood, the LLM identities are anonymized so that the LLM can't play favorites when judging their outputs. The LLM is asked to rank them in accuracy and insight. Stage 3: Final response. The designated Chairman of the LLM Council takes all of the model's responses and compiles them into a single final answer that is presented to the user.

Native support for Mistral API Keys

I‘d love to see the same native support for Mistral models just like with OpenAI, Google and Antrophic models. I am aware of the „Custom Model“-method, but I‘d prefere to simply add the API key and have all Mistral Models pre-made / integrated in the interface. Thank you

Allow uploading audio

Some of the newer models can listen directly to audio, just ilke you can paste in images.

Custom Header and Footer for Word Generator

Allow customization of a preset header and footer (to stay on-brand) in the Word Generator plugin.

Please Help add instructions on how to add brave api

Please help Add instructions on how to add brave api

Automatic Hiding/Deletion of Example Folder

Description: Every time I log in to TypingMind, the standard example folder and chats are recreated and I have to manually delete them. I would prefer to have an option to permanently hide or automatically delete the example folder/chats after each login, so that my workspace stays clean and only contains my own folders and chats. Benefit: This would improve the user experience for those who prefer a clutter-free workspace and don’t need to see onboarding/example content after initial use. Thank you!

Add GPT Image 1.5

Add support for GPT Image 1.5. It would be nice to make it directly accessible like Gemini 3 Pro Image and by way of a plugin like the current GPT Image plugin, which uses the GPT Image 1 model.

Add prompt caching toggle to main typing ui

I think it would be nice to have a quick access toggle like there is with thinking where you can enable caching in the chat. For most models it cost extra and only for 5 minutes. Being able to disable or set to extended easily would help alot save on cost to. Like most chats i will take over 5 minutes to respond so it would be better to disable it. But then in chats were im going back and forth alot its very useful.

API Request: Models, Agents, Plugins & Knowledge Base

1. Model Configuration & API Key Management We require the ability to manage model configurations and rotate API keys programmatically to meet security and compliance standards (e.g., ISO 27001). Currently, API v1.2.0 does not support model or API key management. Requested endpoints: GET /v2/models – List configured models POST /v2/models – Add new model configuration PATCH /v2/models/{id} – Update model details (e.g., rotate API keys) DELETE /v2/models/{id} – Remove model configuration Business justification: Automated API key rotation is critical for enterprise security compliance. Manual configuration creates operational risk and does not scale. 2. Agent Creation & Management Currently, the API only provides read-only access (GET /v1/ai-characters), which limits automation. Requested endpoints: POST /v2/agents – Create new agent PATCH /v2/agents/{id} – Update agent configuration (system prompt, model, access control, etc.) DELETE /v2/agents/{id} – Remove agent Business justification: This capability is required to support Agent-as-a-Service workflows, allowing us to deploy and manage agents automatically through CI/CD pipelines without manual Admin Panel interaction. 3. Plugin Installation & Configuration We need to standardize plugin deployment across multiple TypingMind instances to ensure consistent functionality. Requested endpoints: GET /v2/plugins – List installed plugins POST /v2/plugins – Install and configure plugins PATCH /v2/plugins/{id} – Update plugin settings DELETE /v2/plugins/{id} – Remove plugins Business justification: Lack of programmatic plugin management prevents Infrastructure-as-Code workflows, introduces configuration drift, and increases operational overhead. 4. Knowledge Base Management Currently, uploading and managing knowledge documents is a manual process. Requested endpoints: GET /v2/knowledge – List knowledge bases and documents POST /v2/knowledge – Upload documents and create knowledge base collections DELETE /v2/knowledge/{id} – Remove documents or collections Business justification: Automated knowledge ingestion is essential for scalable enterprise onboarding. Manual uploads do not scale for high-volume or multi-tenant deployments.

Ability to delete from cloud while keeping local storage.

Like stated theirs no ability to delete a image or chat from the could sync while still keeping the local copy. So if i have a large image and want to delete it from the cloud as im not going to use that chat on other devices it deletes it from the local version to.

Ability to set temperature different when thinking is enabled

Some models like claude require temperature to be set at 1 when thinking is enabled and will thro errors if set to anything else. Alot of models preform worse with low temp and thinking on so having the ability to give a separate temperature for thinking vs non thinking would help alot. Currently if you have a custom temp set and thinking enabled you have to change the settings just to even use claude and likely some other models.

Feature suggestion: Project-wide sharing and referencing of chat content within a folder

Description: Please enable chats within a project/folder to access the content (or marked summaries) of other chats in the same project. This would allow, for example, key insights, notes, or results from different threads to be visible and usable for all colleagues/agents in the project—without the need for time-consuming copy & paste or manual uploading of files/markdowns. It would be desirable to have: The ability to mark selected chat messages (e.g., summaries or important steps) as "project knowledge." This marked content is available in other chats in the same project as contextual information (e.g., via search function or automatic integration into system instructions/dynamic prompts). Optional: Project-wide search function across all chats and/or display of selected "shared notes" or "project summaries." Added value: This makes it easier to work on larger projects divided into sub-threads, saves repetition, reduces errors due to loss of information, and speeds up knowledge transfer within the team.

Default Browser when Opening Links

When using as a web app, allow the user to select which browser links are opened in. For example, using the Chrome web app will open any links in Firefox. Similarly, using a Safari web app will open with the default browser in iOS or the one selected in options.