harshithmul

Also thinking deeper, I think your question also is very broad on aspects to answer. Technically in times where there is a gathering of ai and people I think a single chat surface is not the right experience but yeah having a UI saying agents as tabs felt tooo boring and actually fed up.

Just wanted to keep it fun but at the same time can we elevate the experience when there is a common surface where a bunch of ai-npcs and people can meet.

I hope it answers, happy to dig more in a different direction if you felt I understood it wrong

yeah a question with which I started :)

I definitely love the format, but the real reason is each town comes with it's own creative story, theme and interaction rhythm. I first thought about making one town that held them all which I wanted to have for myself or say my own town. But then later thought different people want to have their own representation of it, since it's english, different themes and all about creativity it has no bounds.

In the hindsight I also wanted to figure out which experiences it improves vs which it decorates. Curious what you felt when you experienced it.

hey thanks for the feedback, agreed there are a lot to get right

1. If you have self hosted it yes then we need OpenAI api key, we are also working to get things working with llama so should be available soon (but if from cloud you shouldn't have got this error)

2. We also have moved away from openAI embedding (BGE-M3) to our own hosted in the cloud.

In self host we already have claude code support happy to help there

We faced the same challenge while building SOL (https://github.com/RedPlanetHQ/sol) — a personal assistant that relies heavily on memory for context and continuity.

Getting LLMs to invoke memory tools at the right time is definitely trickier than just wiring up MCP correctly. We're still refining it, but we've made good progress by explicitly guiding the assistant within the system prompt on when and how to use memory.

You can see an example of how we structure this in SOL here: Prompt instructions for memory usage (https://github.com/RedPlanetHQ/sol/blob/964ed23c885910e040bd...)

Using something on similar lines as rules in claude/cursor etc has been working better. It’s not perfect yet, but this combination of prompt engineering and structured tool exposure has been moving us in the right direction.

ps - one of the authors of CORE

We are working on it, LLaMA models are not working good when it comes to finding facts from the message. We should have this out in the next couple of days

Hey thank you, you can find this here https://github.com/RedPlanetHQ/core?tab=readme-ov-file#core-..., it's a docker compose based setup

I was working on something similar https://github.com/harshithmullapudi/tsukuru, so to create a personal website by just writing mdx files. Here is my website https://github.com/harshithmullapudi/harshith.sh

Could you help with how the workflow looked like? How does it ensure that the code is tested?

Ahh, yeah paper trail makes sense. But the ownership aspect of where it went wrong can we understand without the workflow restrictions also right?

Hey, we are still early and are figuring out things. Thanks for being candid, was there something you are interested in seeing in the product? (use case, feature etc)

Current AI functionalities: 1. AI-generated Titles 2. Smart Delegation 3. Duplicate Detection 4. AI Summarization 5. AI filtering 6. Automated Triaging

We are almost done working in 2 other areas 1. AI assistant while create a new issue ensuring the issue has enough information 2. Chat assistant to interact with the tool

Really appreciate the thoughts. The places where we believe AI matters a lot is

1. Ensuring ticket has enough information 2. Offloading some amount task before someone picks it.

These are the 2 key areas we are focusing and we will launch some features around these in the coming weeks. Would love to hear if there are some other places you feel we can focus on?

We tried different approaches, we made a recent change where we started using cohere for re-ranking and embedding which did give really good results.

Happy to create an account for you to try it out. harshith [@] tegon.ai

Thanks for the feedback. We started this 4 months back and these are a few features we build. In the journey and as customers started using we figured out more strong features.

1. Context enricher - To ensure the ticket has enough information 2. AI agents - To offload some part of the work.

We will be getting these out soon

Hey, we have something in personal task management in our roadmap. We are still exploring how to solve that problem. Happy to hear thoughts on how you do it right now. https://github.com/tegonhq/tegon/issues/131

We started seeing people creating random issues we will get it back up after cleaning it a bit

We are still working/improving on the documentation. We use mintlify for documentation https://docs.tegon.ai/introduction.

Noted. We will work on it.

thanks for the catch. We will remove those.

Yes, we use Supertokens and we support email/password and google auth. Google auth is something we added recently as that was requested.

Thanks a ton for the feedback.

1. We started our documentation at https://docs.tegon.ai, though it is not ready yet points taken, and we will work towards making more documentation.

2. We are planning internally to bring more systems to how we manage PRs and to have the right information on them.

3. We love opensource and we have built all our previous products opensource. Keeping our love aside we believe the building of a marketplace for agents/integrations is where the community can help us a ton. We are currently working on making that more modular so that people can start contributing.

What we have done till now started with the basic Project management and there is a lot more to be built and to solve.

Hey, that's where we bring in triage grouping and duplicate identification. We show you what are the similar issues and we are also working on smart merging so to merge the information from both the issues.

My bad. We focused on the performance on the front end currently, as said before we loaded all the data locally and kept the data synced in the background to make the interactions fast. We chose nodejs as that was easy for us to get started with. But definitely, if things get hampered we will start looking into it.

We were experimenting if it could potentially help in taking that off the plate. Feedback taken we will work on improving that or taking a call based on how it performs.

We are also working on a more chat-based approach to interact with the tool. To let you directly handle things from the messaging tools like slack.

True performance is our priority. From the very beginning, we focused on ensuring that our app ran smoothly and quickly. To achieve this, we load all the data locally, allowing you to experience everything at lightning speed. Our background server keeps your data synchronized seamlessly between the server and the client, so you get the best of both worlds: speed and reliability.

We thought auto-generated titles could make life a bit easier. Creating the perfect title for every task can be tricky, especially when things get complicated. So, we added this feature to help with that. (This is also something I faced as a challenge while I was working in Airbyte the community does a great job at explaining the issues but we always have to edit the title to make it more crisp.)

Right now, we’re using the product ourselves and trying it out with a few customers. We’ve noticed that for simple placeholder tasks, a straightforward title works just fine. But for more detailed tasks with longer descriptions, the auto-generated titles seem to work better. We’re still experimenting and gathering feedback to make it even better.

Examples from our board:

Description: - utils.py 228 : Failed to call LLM with the following error: BedrockException Invalid Authentication - An error occurred (UnrecognizedClientException) when calling the InvokeModel operation: The security token included in the request is invalid.

Title: Feat: Investigate Salesforce cloud version error: Invalid Authentication

Description: Triage ux bug - once i decline or accept a triage request, it re-directs me to issues screen. Ideal behaviour to be at the triage list view where you either show removing it from triage or highlighting that an action have been taken on this request.

Title: Redirect user to Triage List after Accepting/Declining Request

Understood, Thank you. Really appreciate all that information.

That sounds exactly like something which we will be launching next week. We are working on an AI assist which will ask questions while creating issues in aspects like

1. If that can be broken into sub-issues

2. If it lacks information according to a template that's created in association with the label. For example: Label bug: You can explain in the template what is expected (Some deployment information, logs, screenshots etc.) and the AI assistant will help in ensuring those contents exist.

We are also exploring if there are more such cases where AI can help. Do let us know your feedback on this once we launch

We are in the process of building those scripts. As we are working with companies for migration we started to improve these scripts to cover more edge cases. These should be out sometime soon.

Otherwise, we have a public API to do CRUD operations for all important entities (issues, labels etc). We are working to get an openAPI spec once that is out we should have all the APIs added to our https://docs.tegon.ai

We haven't yet modularized the AI features of our product. Currently, it's not possible to use the AI features on top of JIRA/GitLab. Our AI capabilities have achieved better accuracy and output because we have more control over the infrastructure and the product's foundations.

Additionally, incorporating agents and managing their workflows require a distinct set of metadata and separate workflows, which we are still in the process of exploring.

Currently, our cloud service isn't publicly accessible, so we've been exploring the best way to demonstrate our product's features and functionality. To address this, we've created some demo credentials and a product instance for demo purposes. Additionally, we've taken a snapshot of a cleaner version of the product, which we refresh weekly.

We will also explore more to see if there is a better way to do this.