One AI layer.
Every app, routed.
NeuralGrid is the AI command center of EternalEngine — the routing layer every other app calls when it needs intelligence. One request, matched to the right Albedo model — Spark for speed, Prime for balance, Titan and beyond for heavy reasoning — by quality, cost, and latency. Automatic fallback if a model stumbles. Every token drawn from one shared wallet you can see.
One request to the right model.
You don't pick the provider — you state the task and the constraints. NeuralGrid does the scoring, the failover, and the accounting.
Describe the task
Send a chat, generation, summarize, extract, classify, or translate request. Add optional constraints — max cost per call, minimum quality, latency ceiling.
NeuralGrid picks the model
A weighted score across quality, cost, and latency selects the right provider for that exact task. Your constraints filter the field before scoring runs.
Fallback keeps it answering
If the chosen provider is degraded, a circuit breaker trips and the request rolls to the next model in the chain. No manual retry, no dropped call.
Logged + costed per tenant
Every call records its model, tokens, and spend against your account. The stats endpoint shows exactly where your AI budget went.
The AI plumbing, done right.
A single multi-provider layer for every text-based AI operation on the platform — chat, generation, summarization, extraction, classification, translation, moderation, and embeddings — with the routing, resilience, and cost discipline built in.
Cost-Based Model Selection
A weighted score — quality 50%, cost 30%, latency 20% — picks the right model per request. Pass maxCost, minQuality, and maxLatency to filter the field before scoring runs.
Fallback Chains + Circuit Breakers
When a provider degrades, a per-provider circuit breaker trips and the request rolls to the next model in the chain. Failover is automatic — your app never sees the outage.
Per-Tenant Cost Tracking
Every call records its model, input and output tokens, and dollar spend against your account. The stats endpoint rolls it up so you always know where your AI budget went.
High-Level Task Operations
Beyond raw chat: summarize, extract structured JSON, classify, translate, moderate, and generate embeddings — each a single endpoint that routes to the model best suited to that job.
Tenant-Isolated, Privacy-First
Every request is tenant-isolated and no PII is ever logged in prompts. AI learning is consent-gated per call, so intelligence never comes at the cost of your data boundary.
The Brains Behind Every App
WikiShelf summarizes through NeuralGrid. ClientForge scores leads, BidForge prices proposals, PostFrame drafts email — all routed here. One AI layer, every app smarter, no per-app provider keys to wrangle.
Infrastructure, not a wrapper.
Every number below is product, not marketing.
Albedo models — Spark, Prime, Titan, Titan+, Colossus, Apex
Quality / cost / latency routing weights
High-level task operations as endpoints
PII fields logged in prompts
Included on every paid plan
NeuralGrid is the AI backbone of EternalEngine — it ships on Team ($49/mo), Pro ($99/mo), and Business ($199/mo). Every app that calls for intelligence routes through it on every paid tier, so the apps you use get smarter the moment you sign up.
See plan comparisonBetter together. By design.
NeuralGrid is the AI layer the rest of the platform calls: WikiShelf summarizes articles through it, ClientForge scores leads, BidForge prices proposals, and PostFrame drafts email — one routing brain, every app smarter.
Common questions, straight answers.
Which plans include NeuralGrid?
Team, Pro and Business. Free and Basic still have the Albedo assistant; NeuralGrid adds the command center that shows and routes the models behind it.
What does it show?
Every active AI feature and the model handling it, plus model cards from quick answers to heavyweight analysis with the tier each belongs to.
Does it change what the AI can do?
It lets you choose which model handles which job within your plan’s ceiling, and it shows usage against your monthly AI allowance.
Run your business from one place.
Five plans, seventeen launch apps, one dashboard. Set up in under 5 minutes — change tiers any time.
From $0 · Basic $9.95/mo · Save 20% on annual billing · Upgrade or downgrade any time