Scalable Self-Hosted AI Without Complexity
While self-hosting AI models can adress intellectual property concerns, traditionally it required specialized MLOps expertise. Codesphere makes it accessible for any developer.
No DevOps needed
GDPR Compliant
Easily Scalable
Fast time2market
With ILTIS GmbH
ILTIS consults enterprises and communal facilities on process optimization.
Partnering with Codesphere allowed the rapid development of new applications to optimize their
workflows.
From PoC to Cross Customer Rollout in < 6 Weeks
Iltis partnered with Codesphere to implement and deploy a speech recognition and
knowledge retrieval use case based on the latest open source AI technologies.
Empowered by Codesphere's ease of use and ability to scale, this PoC project is now used
as a blueprint to scale out the solution across other Iltis customers.
Realtime Speech Recognition
Semantic Search
ERP Integration
Operator Cockpit UI
6 Weeks
Total implementation time
2000+
Documents embedded for Vector DB
100%
GDPR compliant
Composable Architecture
The application consists of the transcription backend, frontend and sentence transformer which creates numerical representations of the transcribed data (embeddings) to store in and query a vector database. Everything hosted on Codesphere.
Sophisticated Ingestion
To make ILTIS’ knowledge available to their experts, we pulled all of their data from their ERP and ran it through a sophisticated ingestion pipeline that converts everything into the right format, chunks the data accordingly and uploads everything to the database. Everything happens fully automated.
Alexander Ott
Geschäftsführer @ ILTIS
Together with Codesphere, we developed a working PoC in just 6 weeks, which allowed us to test and gather feedback incredibly quickly.
Open Source Stack
Frontend
The frontend runs a basic Sveltekit application, taking care of recording the microphone input and sending it as a .wav to the transcription server.
Transcription Server
A Whisper.cpp server running on a Codesphere Pro plan, enabling real time CPU inference for speech to text. No GPU needed, keeping the cost down
Sentence Transformer
A FastAPI server running the sentence-transformer library to create the embeddings needed to store and retrieve ILTIS data from the database
PostgreSQL + pgvector
A PostgreSQL database with the pgvector extension which extends the stored information with vector embeddings and implements a rapid vector search, perfect for RAG use cases
Data Ingestion Pipeline
Node.js server that pulls data from the ERP, parses all PDFs, creates embeddings and checks and fills the .
PDF Server
Sterling PDF endpoint that takes in legacy .doc files and converts them into PDF files that can be parsed by the main Node server.
Freedom in Deployment & Development
The initial structure of the application was developed locally using standard IDEs. Since
Codesphere strips away complex Ops setups and simplifies proxies and path-based routing, making
CORS issues an issue of the past, preview deployments for all stakeholders to see were deployed
easily.
Final tweaks were done right from the Codesphere IDE, stripping away the need for endless git
push and pulls in the final stages of the app development.
Fully GDPR Compliant
Through using an open source and self hosted stack, no data ever leaves the application and its components. Since Codesphere offers servers in Frankfurt, Germany, everything can be hosted within the EU, making it GDPR compliant out of the box.
Get in touch
with Codesphere
Zero-Config
All the cloud provisioning, container orchestration and networking is handled by Codesphere out of the box. You can deploy a running AI model in less than two minutes.
Deploying AI apps on Codesphere is really straightforward.
Helmuth Bederna
Tech Lead @ Medmastery
Reactives: A New Deployment Category
Our patented deployment category, Reactives, surpasses serverless,
buildpacks, Docker PaaS, k8s, and VMs in ease of use and deployment speed, while supporting
enterprise-level complexity.
It excels with high-resource applications like LLMs, simplifying scaling to zero for low-traffic
apps and enhancing efficiency when scaling up by provisioning resources ad-hoc, potentially
reducing infrastructure costs by 50% or more.
Reactive Startup
Eliminate Expertise Bottlenecks
Codesphere takes away the need for dedicated Ops teams, enabling any developer to deploy
performant and efficiently
scaling AI applications without DevOps or MLOps experience.
This eliminates the expensive and time consuming setup & maintenance of
infrastructure as well as traditional challenges associated with scaling.
Learn more or get in touch now!
To get you started right now, we have made some of our training- and informational materials available for you to download below.
Get in touch