The technology behind your product
The AI, web, mobile, backend, and database tools we work with, and why we choose each. We recommend a stack for every project, and we work with yours if you already have one.
How we choose
Start from your constraints
Your existing systems, team skills, data rules, and budget decide the stack before our preferences do.
Prefer proven, open tools
Mature and widely supported technology is easier to hire for and carries less long-term risk.
Test before committing
Models and databases are compared on your own data in a short proof of concept.
AI models
We choose a model for each task by testing it on your own examples, then compare quality, speed, cost, and where your data is allowed to go.
Claude (Anthropic)
Long documents, careful reasoning, codingGPT (OpenAI)
General text, vision, and voiceGemini (Google)
Very long inputs, images, videoLlama, Mistral, Qwen
Open-weight models for private deploymentSmall task-specific models
Classification and extractionAI frameworks and orchestration
The software that connects models to your tools, documents, and approval steps.
LangGraph and LangChain
Multi-step agent workflowsLlamaIndex
Document ingestion and retrievalModel Context Protocol (MCP)
Connecting models to toolsn8n
Light automation between appsPyTorch and Hugging Face
Custom and fine-tuned modelsscikit-learn, XGBoost, LightGBM
Forecasting and scoringSearch and vector databases
How answers are found in your own documents, with the right people seeing the right results.
PostgreSQL with pgvector
Search inside your main databaseQdrant
Large-scale vector searchElasticsearch and OpenSearch
Keyword and hybrid searchPinecone
Managed vector searchSpeech, vision, and documents
For voice agents, camera-based checks, and reading invoices, forms, and scans.
Whisper
Speech to textGoogle, Azure, and Deepgram speech APIs
Live phone callsExotel and Twilio
Phone calls and SMSOpenCV and YOLO
Image and video inspectionPaddleOCR, Tesseract, and vision models
Reading documentsModel serving and MLOps
Running models reliably, and keeping track of what is deployed.
vLLM
Serving open-weight modelsOllama
Pilots and offline useMLflow
Experiment and model trackingDocker and Kubernetes
Packaging and running servicesEvaluation sets and Ragas
Measuring answer qualityWeb frontend
What your customers and staff see in the browser.
React
Interactive web appsNext.js
Websites and apps that rank in searchTypeScript
Typed JavaScriptTailwind CSS
Styling and design systemsVue and Angular
Existing codebasesPlain HTML, CSS, and JavaScript
Fast static sitesMobile and desktop apps
iOS, Android, and desktop, from one codebase or fully native.
React Native and Expo
iOS and Android from one codebaseFlutter
Polished cross-platform interfacesSwift and SwiftUI
Native iOSKotlin and Jetpack Compose
Native AndroidElectron
Windows, macOS, and Linux desktop appsBackend and APIs
The services behind your apps: business logic, integrations, and AI features.
Node.js (NestJS, Express)
APIs and real-time featuresPython (FastAPI, Django)
AI services and data workGo
High-performance servicesJava and Kotlin (Spring Boot)
Enterprise and banking systemsPHP (Laravel)
Business web apps and portals.NET (C#)
Microsoft-based enterprisesREST, GraphQL, gRPC, WebSockets
How systems talk to each otherDatabases and data pipelines
Where your data lives, and how it moves.
PostgreSQL
Main relational databaseMySQL and MariaDB
Existing and web workloadsMongoDB
Flexible documentsRedis
Caching, queues, and sessionsSQLite
Offline and embedded dataClickHouse and BigQuery
Analytics at scaleKafka and RabbitMQ
Event streams and queuesAirflow and dbt
Data pipelinesCloud, DevOps, and monitoring
Hosting, release automation, and keeping an eye on everything in production.
AWS, Google Cloud, and Azure
HostingTerraform
Infrastructure as codeGitHub Actions and GitLab CI
Automated testing and releaseNginx and Cloudflare
Traffic and deliveryPrometheus, Grafana, and Sentry
Monitoring and alertsSecurity and access
Built into every layer, including what the AI is allowed to see.
OAuth 2.0, OpenID Connect, and SSO
Sign-inRole-based access control
PermissionsTLS and encryption at rest, with secrets managers
Protecting dataAudit logs
TraceabilityProgramming languages
We pick the language that fits the job, not the other way round.
| Language | Where we use it | Why |
|---|---|---|
| Python | AI, data, automation, backends | Where the AI and data ecosystem lives. |
| TypeScript and JavaScript | Web frontends, Node.js backends, React Native | One language across web, mobile, and server. |
| Java | Enterprise and banking backends | Mature, strong tooling, and long-term support. |
| Kotlin | Native Android, Spring Boot services | Concise and safe, and the official choice for Android. |
| Swift | Native iOS and macOS | Needed for the deepest Apple platform features. |
| SQL | All data work | The standard way to query and shape data. |
Already have a stack?
We can build on your existing tools. Tell us what you run today and we will suggest where to start.
