Quick Bench
Build a working formula, label, specification, and delivery package without setting up a project.
- Start manually or draft with AI
- Test targets, nutrition, cost, and compliance together
- Deliver directly or move the product into a project
Formulate with food-science AI. Check labels and compliance. Run the full project. Deliver a controlled client package. EdibleIQ connects the work around one product record.
Use the Quick Bench for focused formulation or the Project Workbench for a full product-development program. The product record moves with you.
Build a working formula, label, specification, and delivery package without setting up a project.
Manage one or more products from brief through controlled delivery in a purpose-built product development workspace.
Live nutrition, cost, scaling, targets, and reusable bases.
Supplier documents, multi-file intake, provenance, and searchable data.
Draft, compare, and explain without changing governed work on its own.
FDA-format panels, declarations, evidence, and review findings.
Product workflows, tasks, decisions, files, discussion, and budgets.
Locked versions, exact product facts, integrity records, and client packages.
Upload one file or a complete ingredient folder. EdibleIQ classifies the documents, extracts the useful facts, and keeps every value tied to its source.
Upload a supplier's PDF or a photo. EdibleIQ reads the nutrition panel, matches every line to a standardized USDA nutrient with a confidence score per row β then saves the ingredient, all its nutrient values, and the original document in one click. Traceable to its source, forever.
EdibleIQ AI works with formulas, ingredients, documents, costs, suppliers, projects, and approved knowledge, with the same permissions as the person asking.
The EdibleIQ assistant isn't a chatbot bolted on the side β it's wired into your database, your documents, and every page of the app. It knows your formulas, your ingredients, your costs, your suppliers, and the reference material you've curated.
And the draft doesn't stay in the chat β it lands on your bench as a real, editable formula. Every number live, every total recomputed.
Generate an FDA-format panel from the live formula, then review it through deterministic checks, cited AI analysis, and an artwork-level visual pass.
Required data is enforced. Missing declarations, unresolved ingredients, stale evidence, and incomplete nutrition stop delivery before a version is locked.
Pass, pass with warnings, or fail. Review serving declarations, nutrient math, rounding, %DV, allergens, and required content with citations and next steps.
AI vision checks the rendered label. Typography, rules, spacing, placement, footnotes, and other visual details are reviewed on the label people will see.
EdibleIQ surfaces issues and evidence; your qualified reviewer makes the decision. Findings stay with the exact formula and label version that was reviewed.
Generate a production-ready Nutrition Facts panel straight from your live formula β serving math, FDA rounding rules, and %DV applied per 21 CFR 101.9. Change an ingredient, regenerate: the label always reflects the formula as it stands.
Deterministic, enforced. Every ingredient must carry a label declaration name. Machine-read data a human never verified? Publishing pauses until someone takes responsibility for it β and the acknowledgement is recorded.
AI, against 101.9. Serving-size declaration, nutrient scaling, rounding rules, the full mandatory nutrient set, %DV, allergen statement format β checked against the authoritative reference amount for your product category, resolved server-side.
AI vision, against 101.9(d). The check reads the actual label artwork: font weights and sizes at print, rule weights, %DV placement, thousands separators, footnote format β the things a data check can't see.
EdibleIQ's checks help a qualified reviewer work faster β they don't replace one, and we say so in the product, in the docs, and here. Every check documents exactly what it does and does not verify, with the governing citation. Your regulatory reviewer gets a tool that knows where it stops β which is precisely why they can rely on what it covers.
Build a client-ready package directly from Quick Bench, or coordinate a complex, multi-product delivery in the Project Workbench. Either path freezes the facts that shipped.
Publish is not a status flag. It cuts a numbered, frozen, tamper-evident release: the complete formula snapshot, every document and contract archived, branded Product Specification and Ingredient & Nutrition Sheet PDFs generated, a manifest with an integrity checksum for every file β bundled into one downloadable ZIP.
Run everything from simple task tracking to complex gated workflows. The Project Workbench connects formulas, documents, trials, decisions, approvals, budgets, and client delivery in one product-development workspace.
Every product is a project: trials, client decisions, documents, deadlines. EdibleIQ gives that work a real home β not a generic PM tool bolted beside your formulas, but a project workspace where every task can point at the exact formula, version, and document it's about, under the same roles and audit trail as everything else.
My Work is not a second to-do list somebody has to keep up to date. It is the project workflow, seen from one person's side. When a gate passes, an approval lands, or a task ahead of yours completes, the work it unblocks appears in My Work — already carrying the product, the formula version and the reason it reached you.
EdibleIQ can replace separate project-management and chat tools for product work, while email and calendar integrations extend selected events and approvals into the tools your organization already uses.
The formula in one tool, the plan in a second, the client conversation in a third β and the decision lost somewhere between them. EdibleIQ retires that. Messenger is built into the workspace: real-time team chat where every project has its own channel, your client joins it, and every message lands on the project record where it belongs.
Keep the governed product record in EdibleIQ while messages, dates, files, and approvals reach the everyday tools your team and clients already use.
Gmail and Outlook. Send project updates by email, draft with AI assistance, and bring important client responses back to the product record.
Google Calendar and Microsoft 365 Calendar. Keep trials, decision gates, review dates, and delivery milestones visible on the calendars people follow.
One versioned document repository. Specs, COAs, approvals, and contracts connect to projects, formulas, and tasks, with an in-app viewer and the exact version pinned for review and delivery.
Isolation, controlled access, encryption, safe file handling, and recoverable infrastructure are built into the platform.
Your formulas, documents, and AI context stay inside your customer environment.
Role-based permissions control who can view, edit, approve, publish, or delete.
Data is encrypted in transit and at rest, with protected recovery points and tested restores.
Every upload is quarantined and malware-scanned before use. Unsafe or unscannable files stay blocked.
We break it on purpose. We chaos-test our own stack β kill the database mid-write, kill a worker mid-import, pull services out from under the app β and prove it recovers. A crashed import picks itself back up in about 90 seconds. Nightly encrypted backups run with a dead-man alarm that notices if they ever stop. Deleted records are recoverable, never silently gone.
Every destructive action in EdibleIQ confirms first β a real dialog that names what's about to happen, not a reflex "Are you sure?". Retiring a record is one click, and so is bringing it back. Deleted records land in Recently Deleted, where they can be restored weeks later β not seconds. And a published release is a locked snapshot: nothing edits it after the fact. The undo isn't a support ticket; it's in the product.
Every instance is hosted and operated by us, and that is deliberate. It means the encrypted backups, the restore drills, the monitoring, the security patching and the update path are our responsibility, included in the price β not a checklist handed to whoever on your team drew the short straw. Your data is still yours: export it in portable formats whenever you want. You just never have to run the infrastructure to keep it safe.
Build with supplier materials, USDA references, targets, live nutrition, and costing.
Review nutrition, declarations, labels, evidence, and compliance findings together.
Plan tasks, decisions, documents, approvals, and client collaboration around each product.
Build a locked, versioned package for the client with the exact approved product facts.
One keystroke to any project, formula, ingredient, or company.
Company management wired into projects, ingredients, and quick-add flows.
One versioned home for spec sheets, COAs, approvals, and contracts β linked to formulas, projects, and tasks, with an in-app viewer.
Nothing referenced ever disappears silently β the app tells you exactly what blocks it. And what you do delete lands in Recently Deleted, restorable weeks later.
Your logo on every generated PDF and system email. Your data, your brand.
Standard, simplified, dual-column, tabular, and Supplement Facts β with IU-era vitamin data converted to modern units automatically.
The claims your formula already qualifies for β computed from the declared panel against 21 CFR Subpart D criteria, never guessed.
Drop a folder of spec sheets; the scanning pipeline works through them all β and a crash resumes where it left off, no file lost.
This one came straight from formulators: a system that learns how your experts formulate and helps junior team members level up faster. It's in the product β on the same human-in-control foundation as everything else, with your knowledge staying yours.
When a senior corrects an AI draft β a substitution, a 2% sugar cut, a "that failed at pilot" β EdibleIQ captures the change and the reasoning behind it, so the next person facing a similar brief inherits it instead of relearning it.
Junior formulators get more than a plausible starting point β the assistant draws on how your team solved it before, and what to watch for, from your own reviewed cases.
Nothing teaches the assistant until a senior reviewer approves it. Every captured correction enters a review queue where an expert accepts or rejects it β and a case can be retired just as deliberately. Your experts stay the editors of what the system knows.
Asked for by formulators, built for them β and like everything else here: it proposes, your people decide, and what it learns from your team never leaves your instance.
Development doesn't stop at the label. The November release carries the formula through approvals, costing, trials, and artwork β every decision recorded against the exact version it was made on.
Projects move through named gates β concept, bench, pilot, launch. Each sign-off pins the exact formula, label, and artwork version reviewed, immutably, and publishing waits for the approvals your policy demands.
Cost rolls up from supplier quotes to serving, container, and batch β margin against your target price on the formula screen, and the cost sheet locked into every release snapshot.
Name the targets β cost ceiling, protein floor, a claim to keep β and the optimizer proposes the closest formula. Like every AI feature here: it proposes, your formulator decides.
Label artwork review rounds with pin-point annotations, versioned approvals, and the signed-off proof tied to the release it shipped with.
Every claim you make, where it appears, and its computed 21 CFR eligibility. Change the formula and every affected claim is flagged for re-check on the spot β and publishing is refused while a claimβs check is out of date, so a drifting recipe can never silently break a claim.
Bench trials, sensory panels, and stability studies recorded on the project β the evidence behind every gate decision, next to the formula it tested.
Task boards, timelines and dependencies, comments and attachments on every task β serious project management without leaving the workspace.
Qualify suppliers with documents and scores β and when an ingredient or supplier changes, see every affected formula, label, and claim before it surprises you.
Bring your back catalog in from spreadsheets, send working copies out, and wire up anything else through the public API and webhooks.
Every plan is the same product. Nothing is held back for a higher tier β plans differ only in how many people they cover. Most teams start at 10 users; smaller shops can start at 3 and upgrade mid-cycle, with credit for what they have already paid.
Use a Flex seat for a small gap. Use a pack for the best value.
Move from Studio to Team whenever you need the seats β not at renewal. We credit the unused part of your Studio year against Team and charge the difference for the months that remain, plus a one-time $1,000 upgrade fee. Ten seats cost $15,000 a year however you arrive at them, so growing is never punished with a higher price than starting there. If you already know you will need ten seats inside the first quarter, start at Team β it works out cheaper than upgrading early.
Every plan includes AI under a fair-use limit. If you would rather run without a limit, connect your own AI provider keys and use it as much as you like β your prompts then run on your own account, which is often what a data-governance policy wants anyway. You can switch either way at any time; it is a configuration change, not a new contract. We support the platform either way β with your own keys, the model quota, rate limits and bill are yours.
Every plan includes the full product. Pricing is preliminary and final terms are confirmed in your order form.
Every tier gets the whole product β every feature, AI scanning, assistant & compliance checks, support and updates. What changes between tiers is how many people it's licensed for and the infrastructure and service behind it. Nothing in the software is locked away on a lower tier.
Flat and predictable β one price for the team, billed once a year. No per-seat invoices, no usage bills, no surprise reconciliation.
$1,500 buys you 90 days of the whole product on your own instance. Bring real ingredients, formulate, run compliance checks, publish actual labels β a real development cycle, which is the only honest way to judge it. Decide to stay and the $1,500 is credited toward your first year. Decide not to, and it covered your 90 days. No auto-renewal either way.
And because you're on your own instance from day one, staying is just a licence β the pilot instance becomes your production instance. Nothing migrated, nothing re-imported, not a single formula left behind. Want a look before you commit? A walkthrough is free.
Your tier licenses a number of user accounts, and the app knows its own limit β when you reach it, adding another person asks you to add seats or move up a tier rather than quietly letting the number drift. No per-seat billing, no counting logins, no surprise reconciliation at renewal: bring your whole team up to the number you licensed. Growing mid-year? Adding seats is a conversation, not a penalty.
Need a feature specific to your plant, your workflow, your reports? We scope it with you and quote a flat price per feature β from $2,500 β then build it into your instance. No hourly billing, no open-ended consulting meter.
Everything you put into EdibleIQ β formulas, ingredients, nutrition data, projects, suppliers, and every uploaded document β belongs to you. All of it can be exported in standard, portable formats, ready to import into another database if you ever decide to leave. No lock-in, no hostage data. Need a specific shape β your ERP, a LIMS, a particular schema? We'll build a custom export method for a fixed additional fee.
AI is included in every tier β and your subscription price is your cap.
We never send a usage bill. Bench includes 500 AI actions a month and Scale-Up
1,000 (an action is one document scan or one compliance check β assistant chat
isn't metered, subject to fair use); Enterprise allowances are sized with your contract.
Hit your allowance? Nothing breaks and nothing is billed β AI actions pause until next
month, or add a fixed-price AI pack / step up a tier.
Preliminary pricing β prices may change by the November 2026 release.
Annual billing; monthly available at +15%. Need failover on a smaller tier? Ask about
the HA add-on.
Weβll show you the Quick Bench, the Project Workbench, and a complete client delivery. A 90-day pilot is $1,500 and credited in full if you continue.
Request a pilot