AI
Welcome to Citizen.Chat AI
In our vision all technology should be in the hands of the many, not the few.
For that to be successful, depends on two core elements:
Accessibility. The education and transfer of knowledge would organically follow. But first, all need to be open source.
Privacy. We can pretend that the ones that have the monopoly on new technologies like AI, are happy and willingly giving up their tremendence power. But in fact for the many to have the freedom to innovate, join forces and share knowledge, being able to code interact and communicate anonymously is the second pillar of true disruption and human connection.
Every single day great steps are being made in AI algorithms, but also in the realm of self souvernty, decentralisation and encryption of smart contracts. We are on top of all that is released and where-ever we see the possibility to implement Fully Homomorphic Encryption (FHE) encryption for AI, anonymous verification, and decentralised data storage we will push it to the top of our roadmap. If you have some grants that supports Decentralisation and Private AI, get in touch, we love to work with more people to speed up this developments.
Our approach is creating an AI factory, with all open source elements creating the most flexible and up to date AI for end users and organisations.
#Methodology:
Fully extensible
Built on state of the art open source tech stack. Our RAG ingestion, indexing and execution infrastructure can be plugged into a variety of inference and embeddings models and endpoints. This allows for:
Easy plugin and compatibility with existing infra and adaptions
Always work with the latest up to date versions
Sovereign AI
We took the approach to build towards a state of the art open-source AI system.
Efficient computation
Reduced carbon impact using efficiently state of the art open source LLMs, compute embeddings on cpus
Full control over the AI chain
open source code provides clear understanding of the mechanics involved in a model, access to full customization of the model, full private execution (can be local, or on servers).
Sync with open datasets
We developed this tech for open datasets. We believe there is a huge value in streamlining access to open data for citizens, enterprises, institutions and researchers.
Once released, Nexans will be able to sync its private instances with open datasets such as laws, regulations, public open data, news, open social networks etc...
Augmenting your private data with open data can be used to perform regulatory watch, situation analysis, market studies or overview, market sentiment from news aggregation.
AI in peoples hands
CitizenLab designed citizen.chat for the purpose of making it useful and usable by any citizen. Our approach is based on inclusivity, openness and sovereign access to AI and data.
AI arguably being the biggest revolution in tech also makes individuals afraid AI would replace their jobs when it should be promoted and built as a way to empower their human capital by removing the noise of repetitive cognitive tasks. Empower your teams to focus on what matters and automate the rest, by showing the value AI brings to them
Fully extensible
Built on industry standard open source tech. Our RAG ingestion, indexing and execution infrastructure can be plugged into a variety of inference and embeddings models and endpoints.
Can support:
custom models including openAI embeddings and inference endpoints
any kind of auth
python custom endpoints
custom pipelines, agents and tools
other vector stores
Use Citizen.chat with popular tools:
Tech supported: Mistral / Llama (meta) / Openai / HuggingFace / Anthropic / Coherem / Python / Typescript
Built with: Supabase / Unstructured / Postgresql / Next.js / Vercel / Apache Kafka /
Running on: Kubernetes (K8s) / Google Cloud / Azure / AWS
Last updated
