// roadmap

Built in the open.

Here's exactly what's live, what we're building right now, and what's coming next. No vaporware — just honest progress.

Live now
In progress
Planned
Future
Phase 1
Foundation — Live on Shannon Mainnet
Live

POKT Shannon mainnet

Fully staked gateway and supplier on Shannon mainnet. Real on-chain routing.

Live

OpenAI-compatible API

Drop-in replacement for OpenAI. Any app or library that supports OpenAI works with NodeGhost.

Live

Stripe billing

Automated signups, API key generation, and plan management via Stripe.

Live

Home Assistant support

Works with Extended OpenAI Conversation integration. Full local voice pipeline compatible.

Live

Per-plan rate limiting

Request quotas enforced per API key. Fair and transparent usage controls.

Live

Zero request logging

No content of AI requests is ever logged or stored. Architectural privacy guarantee.

Live

Hosted model access

Access to open source models hosted by independent POKT node operators — no model API key required. Routed through the decentralized network.

Live

Network metrics page

Public transparency dashboard showing live on-chain stats — block height, stakes, suppliers.

Live

Custom endpoint registration

Register your own model server as a backend. NodeGhost handles auth and routing — your model does the inference. Available on all plans.

Live

Agent wallet registration

Autonomous AI agents register using a crypto wallet address. AnChain.AI screens for bad actors. No email, no Stripe, no human in the loop.


Phase 2
Customer Experience
In progress

Usage dashboard

Customer-facing dashboard to track request usage, remaining quota, and billing history.

Live

Public launch

Full site launch, open signups, and announcement to the POKT and Home Assistant communities.

Planned

Usage analytics

Detailed request logs for Business plan customers — timestamps, token counts, latency.

Live

Any model, any provider

Register any OpenAI-compatible endpoint — OpenAI, Groq, Anthropic, or your own self-hosted model. NodeGhost routes it privately through POKT. Your model, your choice.

Planned

Streaming responses

Server-sent events streaming for faster perceived response times in chat applications.

Planned

API key management

Customer self-service portal to rotate, revoke, and manage API keys without contacting support.


Phase 3
Decentralized AI Platform
Planned

Ghost Gateway

A custom NodeGhost gateway layer replacing PATH — built specifically for AI inference with native tool calling support, streaming, and per-service routing logic.

Planned

Streaming responses

Server-sent events streaming for faster perceived response times. Each token delivered as it's generated — no waiting for the full response.

Planned

Third-party supplier marketplace

Independent POKT node operators can list their model endpoints through NodeGhost. Customers get more hosted model options. Suppliers earn relay rewards.

Planned

Image generation

Text-to-image inference routed through the POKT decentralized network. Same privacy guarantees, same relay-based pricing.

Planned

Family & team key packs

One subscription, multiple API keys sharing a single request pool. Perfect for households or small teams.

Planned

x402 payment protocol

HTTP-native micropayments for AI inference. Agents pay per request using the x402 standard — no pre-registration, no API keys, fully autonomous.

Planned

USDC pay-per-inference

Autonomous agents pay with USDC and receive an ng- key automatically. No Stripe, no email — fully permissionless crypto-native billing.


Phase 4
Scale & Ecosystem
Future

NodeGhost Connect

Purpose-built Home Assistant AI integration. One-click setup, no manual configuration required.

Future

White-label API

Let developers build products on top of NodeGhost infrastructure under their own brand.

Future

Dedicated infrastructure

Isolated relay capacity for enterprise customers requiring guaranteed throughput.

Future

SLA guarantees

Uptime SLA with contractual commitments for Business and Enterprise customers.

Future

Additional POKT suppliers

Onboard more node operators to the ai-inference service for redundancy and higher throughput.

Future

Enterprise tier

Custom contracts, volume pricing, dedicated support, and compliance documentation.

Want to influence the roadmap?

NodeGhost is built for the community. If there's a feature you need, we want to hear about it.

[email protected]