APIs, microservices and integrations that hold up in production. Go, Node.js and Python stack on Cloud Run, GKE and AWS. Proposal within 1h.
TL;DR
Dribba is a B2B consultancy with offices in Barcelona and Escaldes-Engordany (Andorra). We build production back-end for companies and startups in Spain: REST and GraphQL APIs, microservices, integrations (ERP, CRM, HL7, Land Registry, SAP BTP), event-driven architectures and data pipelines. Main stack Go, Node.js (NestJS, Fastify) and Python (FastAPI) on Google Cloud Run, GKE and AWS. More than 300 projects delivered since 2011 to seed–Series A startups and Fortune 500 corporations in sectors like health, retail, fintech, logistics and insurtech.
Resumen · datos clave
We work with technology leaders who need reliable back-end: APIs that withstand real peaks, integrations with legacy systems (ERP, CRM, banking, health) and architectures that don't collapse when the product grows. We don't sell digital transformation: we deliver code in production, observable and maintainable.
Our edge isn't the list of technologies —any consultancy claims to have Go and Kubernetes—. It's the judgment to choose when to use each thing. We've shipped back-ends on SAP BTP, data pipelines in BigQuery for retail, transactional engines in Go running on Cloud Run and Node.js APIs under GraphQL Federation. The stack decision is made by the engineer who will maintain it, not the salesperson.
We operate as a single team: product, design, back-end and mobile in the same room (or the same Slack). The senior engineer who designs the API is the same one who deploys and monitors it. There's no handoff between teams: there's one team.
Qué incluye
¿Encaja con lo que necesitas?
30 minutos. Sin pitch comercial. La conversación es contigo y con quien va a construirlo. Respuesta en menos de 1 hora.
Respaldo
+300
proyectos
15+
años
Flutter
partner oficial
What we do
Contract-first design (OpenAPI, GraphQL SDL), OAuth2/OIDC authentication, versioning, rate limiting and auto-generated documentation. APIs that integrate frictionlessly with front, mobile and partners.
When to split and when not to. Real bounded contexts, not fragmentation for fashion. Service mesh with Istio or Cloud Service Mesh if the complexity justifies it.
ERP (SAP, Oracle, Microsoft Dynamics), CRM (Salesforce, HubSpot), health (HL7, FHIR), banking (PSD2, Iberpay), public administration (Land Registry, tax agency). Robust connectors with retries, idempotency and traceability.
Pub/Sub on GCP, Kafka, EventBridge or NATS. Sagas, outbox pattern, dead letter queues. Systems that recover on their own when something fails, instead of paging the on-call.
ETL/ELT with Dataflow, Airflow or dbt. Dimensional modeling, partitioning and clustering in BigQuery. Pipelines that cost little to maintain and much less to redo.
OpenTelemetry, Prometheus, Grafana, Cloud Trace. SLOs defined per product, not per system. Blameless postmortems and runbooks the on-call team understands at 3 a.m.
Stack and technical decisions
We don't have a single stack. We have a library of justified technical decisions. What follows is what we typically ship to production in 2026.
How to decide
| Criterio | Consultancy (Dribba) | Senior freelancer | In-house team |
|---|---|---|---|
| Time to start | 1–2 weeks | 1–4 weeks depending on market | 3–9 months (recruiting + onboarding) |
| Hourly cost (Spain 2026 range) | €75–130/h | €55–110/h | Total employee cost: ~€50–90/h |
| Continuity if someone leaves | Replacement covered by the team | High risk: depends on one person | Vacancy open for months |
| Skill coverage | Back-end + DevOps + mobile + design | What that person knows | Whatever you've hired |
| Code ownership | 100% the client's, transferable | 100% the client's | 100% the client's |
| Scaling up / down | Weekly | Monthly | Quarterly |
| Institutional knowledge | Documentation + ADRs + runbooks | Usually stays in one head | High, if there's no turnover |
| Suitable for | Critical, regulated, multi-stack projects or ones with a deadline | A specific improvement, a concrete reinforcement | Core product with a 12+ month horizon |
It's not a binary decision: many clients start with Dribba and build an in-house team in parallel. We make the handover explicit from the start.
Por qué elegirnos
The engineers who sign your proposal are the ones who write the code. They've shipped back-ends with millions of daily requests to production. There's no layer of juniors behind a PowerPoint.
Saber más →Postgres before Mongo. Modular monolith before microservices. We add complexity when the pain justifies it, not when fashion asks for it. That saves months of refactoring.
The code is delivered ready for your team to take over: documentation, runbooks, ADRs, tests, CI/CD and technical onboarding. We don't leave you tied down by lack of knowledge.
Saber más →Preguntas frecuentes
The Spanish market splits between large integrators (Indra, Accenture, Capgemini, NTT Data) geared to corporate transformation, mid-size software factories (Plain Concepts, Apiumhub, Codium) and specialized boutique consultancies like Dribba. Boutiques are the option when technical quality matters over headcount: small teams, senior engineers who touch the code, and projects with real criticality. Dribba operates from Barcelona and Andorra with more than 300 projects delivered to startups, scaleups and Fortune 500 corporations.
The typical range for an enterprise back-end project in 2026 goes from €25,000 (API + simple integrations, 6–8 weeks with 1–2 engineers) up to €250,000 or more (multi-service platform with CI/CD, observability and a data pipeline). Specialized consultancies' hourly rate in Spain moves between €75 and €130/h. At Dribba the proposal is always at a fixed price or a closed budget per sprint, not by open hours.
There are three tiers. Enterprise mainstream: Java (Spring Boot), .NET and Node.js — the bulk of banking, insurance and public administration. Modern cloud-native: Go, Python (FastAPI), Node.js (NestJS) on Google Cloud Run, GKE or AWS Lambda. And specialized: Rust for critical systems, Elixir/Phoenix for real-time. At Dribba we work mainly with Go, Node.js and Python on GCP, and go into AWS when the client is already there.
Three practical criteria: (1) Timeline — if you need code in production in under 6 months, a consultancy starts in 1–2 weeks; an in-house team takes 3–9 to be operational. (2) Specialization — if the project needs 3 disciplines (back-end, DevOps, mobile) and there's budget for only 1 hire, the agency pays off. (3) Horizon — if the product will live 5+ years with constant evolution, in-house wins long-term. The most common is a hybrid model: an agency to accelerate phase 0–1, in-house for phase 2–N.
For serious enterprise back-end, at minimum: Google Cloud Professional Cloud Architect or AWS Solutions Architect Associate in at least one lead engineer. For regulated projects, add ISO 27001 at the company level and, if it touches personal data, documented GDPR (DPO, records of processing, standard clauses). For banking, SOC 2 Type II. Dribba operates under policies aligned with GDPR and ISO 27001 and works with clients in banking, health and telco under NDA.
Large integrators almost never take projects under €250,000, which pushes them out of the startup market. Boutique consultancies like Dribba work with both profiles: established companies that need a new back-end or to modernize legacy, and seed-Series A startups that want to skip the learning cost. The hourly price is the same; the scope changes.
Node.js: high time-to-market, huge ecosystem, ideal for product APIs with many external integrators and teams already coming from front-end. Go: performance and concurrency for critical services, gateways, transactional engines and anything behind a strict SLA; minimal footprint in containers. Python: the outright winner when there's AI, ML or data pipelines in the domain; FastAPI matches Node in performance for many cases. In practice, serious projects combine two: Node or Python in the product layer, Go in the critical layer.
Sprint 0 in 1–2 weeks from signing. In that phase we deliver: proposed architecture, minimal infrastructure deployed (Terraform on GitHub, CI/CD operational, dev/staging environments), an initial backlog and a release plan. From Sprint 1 there's code in the client repository and biweekly demos. The first functional version in production is usually between week 6 and 12 depending on scope.
¿Listo para empezar?