LLM selection
m
mansona
AI space is every evolving. Would love to have more insight into what LLM is currently being used. Also what emedding model is being used on the inference side as well. What's good today may not be good tomorrow. Are you all using open ai? I'd love the option of using open router even. or better yet give folks a discounted rate, but let them use their own API key.
Log In
Sankari Nair
Hello everyone,
I'm keeping this open, although you can already select from the top frontier LLM models in Recall today
- Claude
- Gemini
- OpenAI
- DeepSea
- Grok and more
That is exclusive to our Max plan: https://www.recall.it/pricing
We've had to do it that way given the very high AI costs associated with these models.
Please note that we also have an API and an MCP so you can connect Recall to the AI model of your choice: https://docs.recall.it/developer/mcp
I'll keep this open for now, as I see that a part of this request is to have your own API key, and I will come back to you when that is something we resolve.
P
Paul Richards
Merged in a post:
Chat gpt
R
Raphael Pinho
Be Able to connect ChatGPT Five. Pay extra for that
P
Paul Richards
Merged in a post:
connect recall ai
P
Paula Paleoca
interrogate with other LLM models (Claude, ChatGPT, etc.) through Docker AI. When will this be possible? Thanks
P
Paul Richards
Merged in a post:
BYOK AI Integration
N
Niks
AI Integration that supports bringing your own API Key & API Host(Openrouter, Groq etc..)
P
Paul Richards
Merged in a post:
Switch to different AI model
C
Curtis Hibbs
I'd like to have the option to choose the AI model that I want to use and/or be able to bring my own API key for the model of my choice.
J
Jay Hightower
I agree with this 100%.
I really love what you’ve built with Recall. From a business perspective, I completely understand why you reserve frontier models for your higher-tier subscriptions. However, for power users who already maintain a high-tier subscription to a service like Claude, having to pay an additional recurring fee to access those same models in Recall creates a double-pay barrier.
If you allowed a Bring Your Own API Key option for these power users, it would solve this for us while also reducing your compute/LLM overhead for high-volume users. We get to leverage the subscription we’re already paying for, and you get to focus on the product-layer value you’ve built, which is the real reason I use Recall.
A
Andrew Shebanow
To mention a few specific models: Qwen, GLM, and Minimax. Supporting an aggregator like OpenRouter or Z.ai would probably be the best way to do this.
r
recall.it.satisfy809
100%.
t
tc
This would be a huge time-saver and avoid copy/pasting back and forth from Recall to another model. I use TypingMind that uses this bring-your-own API concept. It would also save Recall money if people where paying for their own usage (I could imagine).
m
mfaisalk
We should have an option to select LLM Model. Lets say I am not statisfied with ChatGPT or other LLM outputs, I can switch the model for different outputs.
Load More
→