We are Ovia Inc, a studio and lab working in context engineering, AI agent systems, and virtual production — internationally, in English, French, and Spanish.
Our practice: durable authority in AI operations. We design the systems around models — the context, memory, tooling, and review loops that decide whether an agent deployment works in production or only in a demo — and we build media with the same engineering discipline.
This organization holds our own projects beside a curated collection of forks — the open-source shelf the studio evaluates and builds with. Every fork carries an Ovia section at the top of its README: what the project is, why it sits in the collection, and full credit to its upstream.
The shelves:
- Agents & harnesses — agent runtimes, orchestration, and skills: LangGraph, agency-agents, agents-cli, Antigravity, ECC, superpowers
- Voice — realtime voice-agent frameworks: LiveKit Agents, Pipecat
- Memory & context — persistent agent memory and shared context layers: Mem0, Letta, Memorilla, Trail Code
- Models & infrastructure — local runtimes and model tooling: Ollama, airllm, MiniCPM-V
- Production — virtual production and code-first media: cinecode, OpenMontage
- Security & data — security tooling, business intelligence, and data workflows: our own data-science, machine-learning-engineering, and artificial-intelligence repositories
- ai-services — the main site: services, seasonal menus, merchandise, and the OVIA Magazine
- thewebsite — the supplemental site
- token_reducer — token-reduction utilities for context budgets
- differential-calculator — mortgage prepayment penalty calculation (three months' interest and IRD)
- Main site: https://oviainc.github.io/ai-services/
- The Magazine: quarterly, one issue at a time — linked from the main site
#toyoursatisfaction