Product

FeedNova is a cloud formulation platform for aquaculture and livestock nutrition teams. It combines industrial-grade linear optimization (Google OR-Tools) with a curated ingredient database, species-specific constraints, and optional AI workflows, so you move from idea to validated least-cost recipe in one place.

Whether you run a single mill or support clients across regions, FeedNova is built for traceability: every solve is tied to a dataset version, explicit bounds, and exportable results your QA and purchasing teams can trust.

Optimization engine

Least-cost optimization with full nutrient, inclusion and digestible amino acid constraints
380+ ingredients with proximates, minerals, amino acids and environmental proxies (dataset versioning)
Multi-species templates: fish, shrimp, poultry, swine with life-stage aware rules
Scenario comparison, price & volatility views, and ingredient substitution / re-optimization
AI-assisted "Generate Recipe" plus optional RAG knowledge base on your own documents
Organizations, roles, shared credit pools and usage analytics for B2B teams

more than 380+ ingredients

Platform capabilities

Solver engine (OR-Tools)

The core is a linear programming engine that minimizes formula cost subject to linear constraints: crude protein, fat, energy, fiber, ash, minerals, digestible amino acids, ingredient minimums and maximums, and group-level rules. Dual values (shadow prices) help you see which constraints are binding and where the next euro of relaxation would go.

Explicit bounds per nutrient and per ingredient � no black box
Supports multiple solves per session (what-if and substitution flows)
Results: inclusion %, nutrient contribution, cost per kg and per ton

Ingredient database & versioning

Ingredients ship in versioned datasets (e.g. DatasetV1.2) so formulations remain reproducible when you audit or re-run months later. Each record carries proximate analysis, key amino acids, minerals, and optional sustainability proxies for reporting.

Version tags on solves for regulatory and internal QA
Explorer and filters by group, nutrient profile and cost band
Room for your own overrides where your lab data differs from defaults

Species & life-stage constraints

Templates encode industry-aligned rules for major aquaculture species (tilapia, salmon, shrimp, seabass, trout, carp, catfish, etc.) and terrestrial species where enabled. Life stages (juvenile, grower, finisher) adjust nutrient targets and practical inclusion bands so outputs are farm-ready, not just mathematically feasible.

Stage-specific protein, energy and amino acid corridors
Marine ingredient caps and alternative protein strategies (policy-driven)
Easy to compare the same brief across species or stages

Prices, volatility & scenarios

Link market reference prices to ingredients and monitor volatility for purchasing conversations. Run the same formulation against two price sheets, or compare two completed recipes side by side when fishmeal moves � ideal for monthly buying committees and risk reviews.

Historical awareness for "what changed since last month?"
Scenario comparison UI for side-by-side metrics
Supports sensitivity to major cost drivers (oils, meals, grains)

Substitution & re-optimization

When an ingredient is unavailable or too expensive, model a replacement and re-solve. The platform keeps constraint feasibility in view so you do not hand-edit spreadsheets that quietly break nutrient promises.

Swap one or more ingredients and compare cost & nutrient deltas
Preserves dataset version integrity across trials
Faster response to supply shocks or quota changes

AI: Generate Recipe

Describe the diet you need in natural language � species, stage, cost ceiling, ingredient preferences � and the assistant turns it into structured parameters for the solver pipeline. It accelerates briefs; the solver still enforces hard constraints.

Natural-language intake for non-expert stakeholders
Grounded in your tenant configuration and credits policy
Designed as an assistant, not a replacement for nutrition science

Knowledge base (RAG)

Upload PDFs, internal specs, or regulatory excerpts and query them in context. Useful for answering "what did we commit to Customer X?" or "what does our salmon policy say about antioxidants?" alongside formulation work.

Tenant-scoped document stores
Citations to source passages where the model answers
Complements � does not replace � your solver outputs

Organizations, credits & governance

B2B workspaces group users under an organization with a shared credit pool for solver runs and AI usage. Role separation keeps formulation, admin and read-only access clear � especially when external nutritionists collaborate with mill staff.

Org-level subscription and monthly credit envelopes
Usage dashboards for accountability
SuperAdmin tooling for platform operators (separate deployment concern)

Security & deployment posture

The Backoffice runs as a separate authenticated application (typical pattern: public marketing site + app subdomain). Sessions use cookie authentication; DataProtection keys should be persisted in production so logins survive restarts. Pair with HTTPS termination at your reverse proxy.

HTTPS and forwarded headers friendly
Designed for standard ASP.NET Core hosting (IIS, Linux+nginx, containers)
No requirement to embed customer data in third-party LLM prompts beyond your chosen AI policy