I know I have been silly posting, but I assure you that I'm maintaining the work ethic of a Keebler elf. With any luck, I will attain the rank of Santa's elf soon.

Nice!! And indeed, that is the true struggle of local models. I wish you luck. 🙏

Oh, I was replying to with a small salute for pinging me. 😂

o7

This is how a GM feels as of late. https://m.youtube.com/watch?v=xQplfkGF4_A

Howdy! Are you using something like Ollama? You should be able to select 'OpenAI compatible' as an option for 'stacks configure' and point an OpenAI-like endpoint to it. In Ollama's case, this should be http://localhost:11434/v1/ More information on this can be found here: https://github.com/ollama/ollama/blob/main/docs/openai.md Let me know if you have any questions!

This goes into a bit into pioneer territory, that is to say that integrating this functionality outside of a functional template would likely prove costly. At the core, you would need a library like ai-sdk, an OpenAI compatible endpoint, and the ability for a user to provide their own AI API key via UI settings (or provide an application-wide API key yourself that powers the search). There are multiple other ways to achieve this as well, but I don't believe any exist in a predictable, #mkstack-like template at this time. My advice would be to wait for mkstack or another interested party to provide a functional template that an agent can easily integrate into your application. 😓

https://m.youtube.com/watch?v=GKkeDqJBlK8

Yes, aside from the link in the post above, you can also now copy it directly from the share modal. You can also directly link to your profile (which shows your created decks) like so: https://surveil.cards/profile/

Look Ma, I'm on TV.

I feel you, that can be frustrating. There a few different changes required to improve the user experience here - the ability to commit and roll back changes, the ability to reuse prompts, summary of the last session if a new one must be created, and so forth. These interactions technically exist via the command line using the `stacks agent` tool and `git` directly, it is just a matter of translating these workflows to the Shakespeare UI. This will happen, but it will take some time to implement and test.

This site is primarily for building decks and collections, but I think having the ability to play them would be nice. Recreating something with game functions like cockatrice (a free way to play magic over the internet) or the Moxfield deck tester would likely be possible.

Good tidings! This evening, I'm announcing Surveil, a platform for building and sharing Magic: The Gathering decks over #nostr: https://surveil.cards Built using #mkstack, powered by Scryfall API, Surveil aims to provide an intuitive deck-building experience and a viable decentralized option for those who enjoy both concepts. Surveil supports: - All formats searchable within Scryfall - Importing #mtg deck lists - Purchasing cards or whole decks (via TCGPlayer) - Printing full card sheet displays of deck lists - EDHREC suggestions for decks - Sharing cards and decks to your nostr social feed ...and much more! If you're building a new deck or collection, please give it a try. Happy building! #grownostr #surveil

Taking inventory, I have 5 Thinkpads - an x220, T420, T400, T41, and X31. There might be also be a T60 within the coffers of my garage, but I dare not venture forth to confirm.

Ah, yes! Surveil is being worked on actively, even as we speak. It's in the alpha state, but let me know if you have any feedback!

I narrowly escaped the hammer, crawling away to live another day.

Thank you for sharing! There are two probable causes for this behavior. There is a bug I'm currently investigating regarding the service provider being restarted and affecting any pending/active sessions, which could be related. The other behavior, which is more so 'expected', is that after interacting with an agent for a long duration, it will reach a limit within its context window. The only way to move forward from that state (at this time) is starting a new session (similar to what you have described). This is a long-term item that I believe is being investigated, and the eventual outcome would ideally be that the user doesn't have to worry or handle any issues related to the context window being hit. That being said, to your note, the actual project code is unchanged even if that immediate session is lost; the quickest way forward is to start your chat with "we were working with X/Y/Z feature, please review our progress in the codebase and continue". In either case, I know that is actively working to improve the overall user experience for Shakespeare. I very much appreciate your feedback at this stage, and am excited to see the things being built right now. I hope you may continue to enjoy building things out using Shakespeare, and we'll continue working to provide the best tool for the job. 🙂

Oh my, this is likely the same issue as with filtering events I'm the feed. If you haven't already, the prompt to try would likely be akin to - "For the media assets page, we need to filter by authorized authors in our query. This query should follow a similar pattern to the one used in the review feed." As for the session/credits issue, thank you for the heads up! How many credits were lost in your session? I'll see if we can remedy this issue.

No problem! As for image previews, this one can likely be prompted for, though it might take a couple of tries. The gist of it would be: "When sharing art to nostr, please add the blossom image link directly in the post content, after the text, before the hashtags." This is exactly what Reviewstr does, so it's a matter of exposing the image link to the text content of your post - the clients will render/embed it accordingly.

Indeed! There are TCGPlayer purchase links for individual cards and decks, as well as the ability to print full decks with accurate card sizes.

I rebased the project on your stock media version. Import the same project from "https://gitlab.com/soapbox-pub/reviewstr-fork.git" and let me know if that looks better!