Computer vision API
deepface.dev
Managed face-analysis API for developers and agents
A managed API for face verification, embeddings and vector comparison, with self-serve developer and autonomous agent integration.
The need
Integrating a face-analysis library means handling deployment, model requests and usage accounting as well as application code. Developers and AI agents needed an accessible hosted interface around the open-source foundation.
What Tech Local built
We built a managed platform on Sefik Ilkin Serengil’s open-source DeepFace project. Tech Local’s product layer includes the developer dashboard, API gateway, model-server infrastructure, documentation, usage visibility and billing. Direct REST endpoints and hosted MCP tools expose face verification, embedding generation and comparison. Machine-readable documentation guides agents through anonymous signup and self-serve key creation, so they can begin an integration without waiting for manual provisioning.
Important workflows
Integrate focused API endpoints
Use REST requests for face verification, embedding generation and comparison, with documented inputs and traceable request identifiers.
Let an agent begin autonomously
Use machine-readable guidance to sign up anonymously, create a REST or dedicated MCP key and call the supported face-analysis tools.
Manage the integration after setup
Use the developer dashboard and documentation to manage credentials, inspect usage and follow billing beyond the first request.
Implementation
- Open-source DeepFace foundation
- Face verification and embeddings
- Vector comparison
- REST and MCP access
- Agent signup and self-serve keys
- Developer documentation
- Usage dashboard and billing
Build stack
- Next.js · React · TypeScript
- Marketing site and developer dashboard
- Supabase
- Authentication, accounts and usage records
- Node.js · Python · DeepFace
- API gateway and model processing
- MCP · REST · Mintlify
- Developer and autonomous agent integration
- Fly.io · Vercel
- API infrastructure and web hosting
Practical value
Developers can integrate face-analysis capabilities without hosting the inference infrastructure themselves. Agents can discover the documentation, create their own account and credentials, and call focused tools. The original DeepFace library remains visible and attributable to its creator, while the hosted platform supplies the surrounding product and operational layer.