Promptly retweeted
Llama 3.1 from @AIatMeta can be run locally using Ollama and LLMStack. 🦙
Instructions in the guide below. 🔽
A quick guide to building AI apps with a locally running Llama3 model in a no-code way 🚀
We use @AIatMeta's Llama3 model with @ollama and @trypromptly's LLMStack as the no-code framework to create a simple recipe generator. 🤖
Promptly retweeted
A quick guide to building AI apps with a locally running Llama3 model in a no-code way 🚀
We use @AIatMeta's Llama3 model with @ollama and @trypromptly's LLMStack as the no-code framework to create a simple recipe generator. 🤖
Promptly retweeted
Added Llama-3-70B support to @trypromptly with @GroqInc. Groq's inference speeds should worry nvidia / openai
Here is a quick app that shows the inference speeds of Groq and compares capabilities of Llama-3-70B with GPT-4 Turbo
Check out the app at trypromptly.com/a/groq-llama…
Promptly retweeted
Exciting news! Mixtral-8x22B and other Mistral AI models are now live on Promptly! 🎉
Test the model in our Playground and get ready to build innovative apps! 💡
We just released Mixtral-8x22B-v0.1 and Mixtral-8x22B-Instruct-v0.1:
- Free to use under Apache 2.0 license
- Outperforms all open models
- Native function calling
- Masters English, French, Italian, German and Spanish.
- Seq_len = 64K
mistral.ai/news/mixtral-8x22…
Promptly retweeted
Join us next week for an exciting webinar where we will demonstrate how to build a LLM Agent with internet access in just minutes. 🎉
Mark your calendars for an exciting webinar on Wednesday, February 14th!
Join us as we demonstrate how to create a no-code stock analyzer agent using LLM function calling, in collaboration with @SingleStoreDB 😃📅📚
Promptly retweeted
Replying to @trypromptly
@trypromptly showed up in today's edition of Superhuman newsletter
thanks for the shout out @heykahn
superhuman.ai/p/neuralink-im….
Promptly retweeted
New AI agents on the list:
🛠️ @Kompas_AI - Pick LLM & build custom agents
🚀 @trypromptly - GenAI for Your Enterprise
🤖 Open-source SW agent: CodeFuse-ChatBot
Check out all products: e2b.dev/ai-agents
(We are getting a high number of submissions - thank you for your patience with adding new products.)
Promptly retweeted
I'm sure you've seen Milei's WEF speech in English by @HeyGen_Official 👀
Did you know HeyGen's Realtime Avatar API can make your avatar repeat any text in realtime?
✨ Excited to share that you can now use it from LLMStack 🚀
Make your avatar answer questions from your documents in realtime without code 🧵
Promptly retweeted
We’ve recently added web browsing capabilities to agents on LLMStack. Here is how it differs from ChatGPT and other platforms:
🔍 Automate workflows with programmable browser
👁️Leverage @OpenAI's gpt-4-vision-preview to automagically browse the internet
📹Streaming video of the browser in action in real time
More in the 🧵
Promptly retweeted
This is our new video for @trypromptly landing page.
📢 We have added a ton on new features to Promptly / LLMStack over the past few months. Stay tuned for a lot of content around agents, RPAs, Custom GPTs etc.,
Promptly retweeted
7/ @trypromptly :
Promptly / LLMStack is an open-source platform excelling in building AI agents and applications with your own data. It supports major model providers and diverse data sources, enabling collaborative app building with detailed permission controls.
Promptly retweeted
We have been building retrieval augmented generation (RAG) pipelines at Promptly for a while and now include them in LLMStack 🚀
Below are some techniques we have come across when trying build effective RAG pipelines in production 📈
🧩 Hybrid Search: Combining semantic & keyword search for better results.
📝 Summaries: Store summarized data to remove noise & reduce token count.
🔀 Overlapping Chunks: Pass overlapping chunks for context-rich generation.
🎯 Fine-tuned Embeddings: Customize embeddings for domain-specific precision.
📚 Metadata Matters: Provide context with metadata for improved understanding.
🔁 Re-ranking: Diversify top results based on metadata & keyword matches.
🤔 Lost in the Middle: Address the "Lost in the Middle" problem with context re-ordering.
Read more about these techniques and about RAGs in general in our blog post at llmstack.ai/blog/retrieval-a…
Promptly retweeted
Check out the project at github.com/trypromptly/LLMSt…
loom.com/share/1399a39c19394… is a quick demo of the platform
Promptly retweeted
Pushed some major changes to LLM Stack 🚀
- Apps are now versioned allowing tracking of changes
- Seamless collaboration on prompts and apps
- YAML to define apps and LLM chains which should help power users build apps more quickly
Promptly retweeted
Empower everyone in your org to leverage generative AI by deploying LLMStack (github.com/trypromptly/LLMSt…) 💪
Check out the quickstart video: loom.com/share/1399a39c19394… 🤖
Promptly retweeted
Replying to @trypromptly
Excited to partner with @googlecloud to bring low-code LLM app development to Google's Vertex AI models 🤝
cloud.google.com/blog/topics….
#GoogleCloudNext
Promptly retweeted
🎇 LLMStack, our open-source low code LLM app platform made it to today's @bensbitesdaily 🎇
Check out the project at: github.com/trypromptly/LLMSt…
thanks @bentossell 👋
Promptly retweeted
Introducing LLMStack: github.com/TryPromptly/LLMSt… 🚀🚀
an open-source platform to production ready build LLM apps with your data. You can deploy it locally for personal use or on a server to use by your entire organization.
What can you do with LLMStack? 🧵















