Ainfera is a five-layer platform for deploying, discovering, and settling AI agents across frameworks.

1. Trust — five-dimension agent trust scoring (Reliability, Quality, Cost-efficiency, Security, Compliance) grounded in real execution telemetry. Geometric mean aggregation — any dimension at zero collapses the overall score. There is no gaming this.
2. Discovery — W3C DID registry (did:ainfera:agent:*) with cross-framework semantic search and verifiable credentials.
3. Compute — sandboxed execution with sub-500ms kill switches. Docker today, Firecracker microVMs on the roadmap.
4. Billing — Stripe Connect three-way splits. Creator 85% · Platform 5% · Compute 10%. Fiat-native, no blockchain.
5. Orchestration — MCP and A2A v1.0 with policy-enforced routing (Cedar + OPA policy engine arrives in Stage 1).

More repositories open as we expand the platform.

Ainfera is in private beta. Developer alpha is shipping now — the CLI, SDK, and API are live. Platform features roll out through Q2–Q4 2026.

Pre-seed funding round open in Q2 2026. Built on NVIDIA NIM, NeMo Guardrails, and Stripe Connect. Incorporated as a Singapore Pte. Ltd. (post-funding close).

pip install ainfera
ainfera auth login --key ainf_your_api_key
ainfera status
ainfera deploy --demo

See the CLI README for a full command reference, and the API docs for endpoint-level detail.

• NVIDIA Inception & Ignition — application submitted
• SMU School of Computing — research partnership in active discussion with Prof. David Lo (OUB Chair Professor; ACM Fellow; IEEE Fellow; PI on the NRF-funded TrustedSEERs agent trust research program)

• General: hello@ainfera.ai
• Security: security@ainfera.ai
• Careers: careers@ainfera.ai

_{© 2026 Ainfera Pte. Ltd. · Singapore · Apache 2.0 for open-source components}