Flávio Juvenal

- https://docs.llamaindex.ai/en/stable/understanding/putting_it_all_together/apps/fullstack_app_guide/ - https://docs.llamaindex.ai/en/stable/understanding/putting_it_all_together/apps/fullstack_with_delphic/ - https://docs.llamaindex.ai/en/stable/understanding/putting_it_all_together/chatbots/building_a_chatbot/ - https://docs.llamaindex.ai/en/stable/understanding/putting_it_all_together/q_and_a/terms_definitions_tutorial/ What features are in those tutorials that we can provide in a reusable app like ours? Related task: https://github.com/vintasoftware/django-ai-assistant/issues/13

- https://www.copilotkit.ai/ - https://sdk.vercel.ai/docs/ai-sdk-core/overview - https://github.com/phidatahq/phidata - https://github.com/ragapp/ragapp - https://github.com/metaskills/experts - https://github.com/Chainlit/chainlit - https://github.com/khoj-ai/khoj/ - https://markprompt.com/ - https://github.com/julep-ai/julep - https://github.com/dagworks-inc/burr - https://github.com/truefoundry/cognita - https://langroid.github.io/langroid/quick-start/chat-agent-tool/ What features are in those tools...

Support Vision / Image attachments

1

https://platform.openai.com/docs/guides/vision See: https://python.langchain.com/v0.2/docs/how_to/multimodal_inputs/ Facilitate mediafile image attachments too. And add an example.

Rate-limiting

1

Support easy configuration of rate-limiting for our endpoints. Support a global rate-limit, but a per request/user rate-limit checker function too.

https://python.langchain.com/v0.2/docs/how_to/chatbots_memory/#modifying-chat-history https://python.langchain.com/v0.1/docs/expression_language/cookbook/prompt_size/

It's usually helpful to include the current datetime to system prompt (instructions). But there are more common system prompt techniques we can use. See: https://github.com/phidatahq/phidata/blob/e4b9e5a3322fc9879a6d1be11d5f30973fe239a8/phi/assistant/assistant.py#L579-L627

How to make Celery and langgraph work together? Is it worth? What's the best way to coordinate assistants with background work and run it all outside the request-response cycle?

- [ ] Django mailer - [ ] Django model-based RAG - [ ] Django Mediafile-based RAG

- Any SOC2 facilities - Any GDPR facilities - Encryption - Data retention controls - PII removal and/or anonymization - Team management - Fine-grained permissions