Stack · 211 technologies · 11 layers
The stack. Chosen for your system, not for fashion.
Interfaces, services, data, AI, cloud, devices and everything that keeps them running. We work across the whole system, default to proven tools, and bring in new ones only where they earn their place.
211 technologies · 11 layers
L-01 · 17 technologies
Web interfaces
Fast, accessible interfaces in TypeScript, rendered on the server where it helps and animated only where it means something.
Practice: Product Engineering, Product & Experience Design
- TypeScriptTyped JavaScript
- JavaScriptLanguage
- ReactUI library
- Next.jsReact framework
- Vue.jsUI framework
- NuxtVue framework
- AngularApplication framework
- SvelteUI framework
- AstroContent sites
- Tailwind CSSStyling
- shadcn/uiComponent system
- Radix UIAccessible primitives
- ViteBuild tooling
- StorybookComponent workshop
- GSAPMotion
- Three.js3D and WebGL
- D3Data visualization
L-02 · 13 technologies
Mobile & desktop
Native where the platform matters, cross-platform where shared code wins, and desktop apps that feel at home on each operating system.
Practice: Product Engineering
- SwiftiOS and macOS
- iOSPlatform
- macOSPlatform
- KotlinAndroid and services
- Jetpack ComposeAndroid UI
- AndroidPlatform
- React NativeCross-platform apps
- ExpoReact Native tooling
- FlutterCross-platform UI
- DartLanguage
- TauriDesktop apps
- ElectronDesktop apps
- RustSystems language
L-03 · 21 technologies
Backend & APIs
Services with typed contracts, durable jobs and clear boundaries, in the language your team can keep running after we leave.
Practice: Product Engineering, Cloud & Platform Engineering
- Node.jsRuntime
- BunRuntime
- DenoRuntime
- PythonLanguage
- FastAPIPython APIs
- DjangoWeb framework
- GoLanguage
- JavaLanguage
- SpringJava framework
- .NETPlatform
- PHPLanguage
- LaravelPHP framework
- Ruby on RailsWeb framework
- NestJSNode framework
- ExpressNode framework
- HonoEdge-ready APIs
- GraphQLAPI style
- tRPCTyped APIs
- OpenAPIAPI contracts
- gRPCService-to-service RPC
- TemporalDurable workflows
L-04 · 26 technologies
Data
One well-designed relational database first, then caches, search, streams and warehouses when the product actually needs them.
Practice: Product Engineering, Applied AI & Data
- PostgreSQLRelational database
- MySQLRelational database
- MariaDBRelational database
- Microsoft SQL ServerRelational database
- SQLiteEmbedded database
- TursoEdge SQLite
- SupabasePostgres platform
- NeonServerless Postgres
- MongoDBDocument database
- RedisCache and queues
- ClickHouseAnalytics database
- DuckDBIn-process analytics
- ElasticsearchSearch
- OpenSearchSearch
- MeilisearchSearch
- Apache KafkaEvent streaming
- RabbitMQMessaging
- NATSMessaging
- PrismaORM
- DrizzleORM
- SnowflakeData warehouse
- BigQueryData warehouse
- DatabricksLakehouse
- Apache SparkData processing
- Apache AirflowPipeline orchestration
- FirebaseApp backend
L-05 · 21 technologies
AI & machine learning
Hosted and open-weight models behind evaluation, retrieval and guardrails, chosen per task for quality, latency, cost and where your data may go.
Practice: Applied AI & Data
- AnthropicClaude models
- OpenAIGPT models
- Google GeminiGemini models
- Mistral AIOpen and hosted models
- LlamaOpen-weight models
- Hugging FaceModel hub
- OllamaLocal models
- vLLMModel serving
- PyTorchDeep learning
- TensorFlowDeep learning
- scikit-learnClassical ML
- LangChainLLM tooling
- LangGraphAgent workflows
- Model Context ProtocolTool access for agents
- QdrantVector search
- pgvectorVectors in Postgres
- MLflowExperiment tracking
- JupyterNotebooks
- pandasData analysis
- NumPyNumerical computing
- PolarsFast dataframes
L-06 · 26 technologies
Cloud & platform
Managed services and pay-per-use compute by default; clusters and reserved capacity only when the numbers say so. Always in your accounts.
Practice: Cloud & Platform Engineering
- Amazon Web ServicesCloud
- Google CloudCloud
- Microsoft AzureCloud
- CloudflareEdge network
- Cloudflare WorkersEdge compute
- VercelFrontend cloud
- NetlifyFrontend cloud
- Fly.ioApp platform
- RenderApp platform
- RailwayApp platform
- DigitalOceanCloud
- HetznerCloud and servers
- DockerContainers
- KubernetesOrchestration
- HelmKubernetes packages
- TerraformInfrastructure as code
- OpenTofuInfrastructure as code
- PulumiInfrastructure as code
- AnsibleConfiguration
- NGINXWeb server
- CaddyWeb server
- TraefikProxy
- LinuxOperating system
- UbuntuOperating system
- UpstashServerless data
- MinIOObject storage
L-07 · 24 technologies
Delivery & observability
Every change reviewed, tested and scanned on its way to production, and every system instrumented so problems are seen before users report them.
Practice: Cloud & Platform Engineering, Security & Reliability
- GitVersion control
- GitHubCode hosting
- GitHub ActionsCI/CD
- GitLabCode hosting and CI
- BitbucketCode hosting
- JenkinsCI/CD
- Argo CDGitOps delivery
- OpenTelemetryTracing and metrics
- PrometheusMetrics
- GrafanaDashboards
- SentryError tracking
- DatadogMonitoring
- New RelicMonitoring
- PagerDutyOn-call
- JaegerDistributed tracing
- PlaywrightEnd-to-end testing
- VitestUnit testing
- JestUnit testing
- CypressEnd-to-end testing
- k6Load testing
- SonarQubeCode quality
- RenovateDependency updates
- DependabotDependency updates
- ESLintLinting
L-08 · 14 technologies
Security & identity
Standard identity protocols, managed secrets and continuous scanning, so security is part of how the system is built rather than a phase before an audit.
Practice: Security & Reliability
- OWASPSecurity standards
- OpenID ConnectIdentity protocol
- WebAuthnPasskeys
- Auth0Identity platform
- OktaIdentity platform
- KeycloakIdentity server
- ClerkApp authentication
- HashiCorp VaultSecrets management
- 1PasswordTeam secrets
- Let's EncryptCertificates
- SnykDependency scanning
- TrivyContainer scanning
- TailscalePrivate networking
- WireGuardVPN
L-09 · 21 technologies
Devices, IoT & real time
Brokers, pipelines, fleet platforms and companion apps are ours. Firmware and hardware we build alongside your hardware team or specialist partners.
Practice: Connected & Real-time Systems
- MQTTDevice messaging
- Eclipse MosquittoMQTT broker
- EMQXMQTT broker
- AWS IoT CoreDevice cloud
- Azure IoT HubDevice cloud
- Node-REDFlow automation
- InfluxDBTime-series database
- TimescaleDBTime-series Postgres
- Espressif ESP32Wi-Fi and BLE chips
- Nordic SemiconductorBLE chips
- STMicroelectronicsMicrocontrollers
- ArduinoPrototyping
- Raspberry PiEdge computers
- Zephyr RTOSEmbedded OS
- ESPHomeDevice firmware
- Bluetooth LEShort-range radio
- ZigbeeMesh radio
- LoRaWANLong-range radio
- Home AssistantHome automation
- Socket.IOReal-time web
- WebRTCReal-time media
L-10 · 23 technologies
Integrations
Payments, messaging, CRMs and the SaaS tools your business already runs on, connected with contracts, retries and monitoring.
Practice: Product Engineering
- StripePayments
- PayPalPayments
- AdyenPayments
- TwilioSMS and voice
- ResendEmail
- MailgunEmail
- SlackTeam messaging
- JiraIssue tracking
- LinearIssue tracking
- NotionDocs and wikis
- SalesforceCRM
- HubSpotCRM
- ShopifyCommerce
- ZapierAutomation
- n8nAutomation
- AlgoliaHosted search
- MapboxMaps
- Google MapsMaps
- ContentfulHeadless CMS
- SanityHeadless CMS
- StrapiHeadless CMS
- PayloadHeadless CMS
- WordPressCMS
L-11 · 5 technologies
Design & prototyping
Design in the same tokens the code uses, prototypes people can click and motion that is specified, not guessed.
Practice: Product & Experience Design
- FigmaInterface design
- FramerInteractive prototypes
- MiroWorkshops and mapping
- RiveInteractive motion
- LottieFilesLightweight animation
Logos and product names are trademarks of their owners, shown only to name the technologies we work with. No partnership or endorsement is implied. Some brands don’t allow their logos to be used this way, so we list them by name. Logos from Simple Icons (opens in a new tab).
How we choose
The best tool is usually the one you can still run in five years.
“Adding the technology is easy, living with it is hard.”
01
Boring by default
Proven tools have known failure modes and people who can run them. We spend novelty only where it makes your product different.
02
Managed over self-hosted
If a provider runs it well for less than it costs to operate yourself, we use the managed service and keep the time for your product.
03
Standards over lock-in
SQL, OpenAPI, OpenTelemetry, OIDC and containers keep your options open when a vendor changes its prices or its mind.
04
Your team can run it
We favor what your engineers already know, and document what they don’t, so the system never depends on us.
Start a conversation
Running something that isn’t listed?
Tell us your stack. If it runs in production, we’ll tell you honestly how well we know it.