Skip to main content

πŸ’»-coders 2024-12-08

Summary​

The chat focused primarily on configuring and running the openai-compatible model, specifically with .env file adjustments. Michaelben sought guidance for this configuration process while yodamaster726 shared his experience of successfully utilizing a large OLLAMA 39gig model on high RAM MacBook Pro hardware.

FAQ​

  • How to configure with openai-compatible model? Not local ollama, what should I do in .env file? (asked by @michaelben)
  • I have a RTX 3090. How can it be used for testing OLLAMA models? (asked by @youngphlo)
  • What are the benefits of buying a MacBook over building your own PC? What makes it worthwhile for certain users like artists and creatives, but not coders or builders? (asked by [Shisho] (03:15))
  • Why do people buy macbooks despite custom built pcs being cheaper with more hardware options? Is there a specific reason why some prefer the reliability of MacBooks over PC builds, even if they are aware that it's essentially an expensive cell phone processor? (asked by agwnl)
  • How long are the session tokens valid for generally? For Twitter. (asked by [techfren (04:44)])
  • Is OKai able to make a trading bot? (asked by [Ȑ̡͘SΜ·Ν‚Μ‹(05:09)])
  • Can the okai-starter run in WSL terminal? Or does it lack graphical interface support on localhost:3000, resulting in 'Cannot GET / Failed to load resource' error? (asked by [Konstantine (05:08)])
  • In a custom action, how can I access the user handle of the person that triggered an action? Is it within the character file? (asked by fOfer (05:26))
  • Is there a way to get text before making a request and after, but before sending response back to client? This would be useful for database interactions. (asked by [꧁Ninja_Dev꧂] (05:36))
  • On the other hand, if I want to generate messages within plugin is there an easy way currently available? (asked by [fOfer] (05:59))

Who Helped Who​

  • @JamLong | Gaia 🌱 helped @michaelben with Configure OLLAMA models using environment variables by providing Michaelben asked about configuring with openai-compatible model and received guidance on checking .env file.
  • agwnl helped [Shisho] (03:20) with Discussing laptop options by providing Shisho provided advice on considering custom-built laptops for better performance and cost efficiency
  • Shisho helped Grivier with Resolve JSON parsing error by providing Debugging non-JSON characters in response data
  • techfren helped Thanks for the info on VPS pricing. with Provided information about low-cost Linux VPS options and AWS grant. by providing [Shisho (04:10)]
  • Konstantine helped Reassured that Reddit will be handled before PR. with Provided reassurance about handling a task. by providing [AIFlow.ML (04:59)]
  • Ȑ̡͘SΜ·Ν‚Μ‹ helped Advised to use the client for OKai-starter. with Provided guidance on using a specific tool. by providing [AIFlow.ML (05:20)]
  • Provided information about the client folder and its use for building front-end. helped Ȑ̡͘SΜ·Ν‚Μ‹ with [fOfer] (05:26) by providing [AIFlow.ML] (05:39)
  • Confirmed that the agent replies to comments on tweets, but mentioned an issue with testing. helped Bunchu with [yodamaster726] by providing ꧁Ninja_Dev꧄
  • agwnl helped big trav with Configuring OKai agents to post on X platform by providing dotbear89 provided guidance to big trav about setting up Twitter client in agent's character.json and .env file.
  • @Kevin Mok helped @AIFlow.ML with Stuck querying when @'ing it in Discord by providing Help with issue #921 in OKai.

Action Items​

Technical Tasks​

  • Configure .env for openai-compatible model (mentioned by michaelben)
  • Change default character in pnpm start to use ollama 39gig model on macbook pro m4 with 48GB RAM (mentioned by yodamaster726)
  • Consider building a custom laptop for better performance (mentioned by Shisho)
  • Developer should investigate unexpected "<" character causing JSON parsing error (mentioned by Grivier)
  • Apply for an AWS grant (mentioned by [Shisho (04:11)])
  • Improve access to user handle within custom actions (mentioned by [fOfer, Grivier])
  • Enhance pre-request text retrieval and post-response storage in database for user interactions with bot (mentioned by [fOfer, AIFlow.ML])
  • Develop plugin message generation using existing composeContext and generateText methods or create a runtime method for prompt handling (mentioned by [fOfer, Bunchu])
  • Investigate issue with bot not responding to replies in Tweets and optimize if necessary (mentioned by [꧁Ninja_Dev꧂])
  • Add Twitter client configuration in agent's character.json and .env file for X platform posting. (mentioned by agwnl)
  • Resolve Tweetbot error on ttwitter (mentioned by @copycat)
  • Activate Solana plugin for Bird Eye data service API key (mentioned by @st4rgard3n)
  • Investigate plugin parameters issues (mentioned by @DL)

Documentation Needs​

  • Clear session tokens and start a new one after hitting 404 error. (mentioned by [Shisho(04:13)])
  • Review code contribution process for the repository, focusing on replies and active search (mentioned by [sam-developer, Bunchu])
  • Update reaction role bot with new character info and emoji roles. (mentioned by @jin)
  • Activation process for plugins to be clarified in the index setup file. (mentioned by @TheHunter@12:30)

Feature Requests​

  • Explore MLX OKai for running models efficiently despite constant changes and fast model runtimes. (mentioned by AIFlow.ML)
  • Evaluate the benefits of MacBook's reliability and ease-of-use over PC builds (mentioned by agwnl)
  • Implement a feature for OKai agents that allows them to post on X platform. (mentioned by dotbear89)