Tonic pricing: what’s included in Structural vs Textual vs Fabricate, and how do plans differ?

Most teams first come to Tonic with a simple question: “What do we actually get with Structural vs Textual vs Fabricate—and how does pricing scale as we connect more data?” Underneath that is the real tension: you need production-like data for development and AI, but you can’t keep copying raw production into lower environments and RAG pipelines without blowing up your risk profile.

This guide breaks down how Tonic’s products map to that problem, what’s included in each, and how the pricing models differ so you can align the right plan to your workflows—not the other way around.

Quick Answer: Tonic pricing is built around three products—Structural for structured data de-identification and synthesis, Textual for unstructured data privacy and AI prep, and Fabricate for agentic synthetic data generation. Structural is priced primarily by connected source data volume and plan tier; Fabricate and Textual are packaged around usage and workflows for synthetic data generation and unstructured de-identification, with options ranging from pay-as-you-go to enterprise contracts.

The Quick Overview

What It Is: A synthetic data and data de-identification product suite designed to give you high-fidelity, privacy-safe data for software development and AI, without copying sensitive production data into lower environments.
Who It Is For: Engineering, QA, data, and AI teams in privacy-sensitive environments (finance, healthcare, SaaS, public sector) who need realistic test data, safe AI training/rAG pipelines, and governed access to production-shaped data.
Core Problem Solved: You can’t safely ship on top of raw production data, and you can’t reliably test or train on fake, manually masked, or toy datasets. Tonic preserves utility—relationships, behavior, and distributions—while removing sensitive information at scale.

How Tonic’s Products Fit Together

At a high level:

Tonic Structural → Transform existing structured production data into de-identified, high-fidelity datasets for dev, QA, and analytics.
Tonic Textual → Detect, redact, tokenize, and synthesize unstructured text (and documents) ahead of RAG ingestion or LLM training.
Tonic Fabricate → Generate from-scratch synthetic structured and unstructured data via an agentic Data Agent—perfect when you can’t or shouldn’t connect to production at all.

You can deploy one product or all three. Under the hood, they’re designed to align with how teams actually ship:

Hydrate lower environments safely (Structural).
Turn production databases into referentially intact test data—same schemas, same behavior, no real identities.
Prepare unstructured data for AI and search (Textual).
Run NER-powered pipelines on tickets, PDFs, emails, call transcripts, etc., to remove or transform sensitive entities while preserving semantic realism.
Generate net-new datasets and artifacts (Fabricate).
Describe what you need to a Data Agent and get synthetic databases, documents, and mock APIs you can share freely.

Tonic Structural: What’s Included and How Pricing Works

Tonic Structural is the engine for structured/semi-structured data—relational databases, data warehouses, and similar stores. It exists because DIY masking scripts and CSV exports don’t scale, and they tend to break exactly where it hurts: foreign keys, joins, and statistical distributions.

What Tonic Structural Includes

Structural focuses on preserving app behavior and test coverage while removing sensitive information:

High-fidelity de-identification & synthesis
- Column- and table-level transforms for PII/PHI and other sensitive data.
- Mix of deterministic masking, format-preserving encryption, and synthetic data generation.
- Cross-table consistency so the same user, account, or entity is transformed the same way everywhere it appears.
Referential integrity across the dataset
- Maintains foreign key relationships so applications, queries, and reports still work.
- Designed to avoid the classic failure mode where masking breaks joins and test suites.
Subsetting with referential integrity
- Extract a minimal but representative slice of production that still “behaves like” the full dataset.
- Preserve relationships across tables in the subset; don’t strand orphan records.
- Option to license Subsetting-only, in conjunction with Tonic Ephemeral, if your core need is small, safe, on-demand environments.
Schema awareness and governance
- Schema change alerts to prevent new sensitive columns from silently bypassing your privacy rules.
- Custom sensitivity rules so you can encode what counts as sensitive in your environment.
Deployment options for regulated environments
- Tonic Cloud or self-hosted deployments.
- Enterprise-grade governance with SSO/SAML and integrations into CI/CD, plus auditability to keep compliance continuous rather than episodic.

The outcome: production-shaped datasets that are safe to pipe into staging, QA, and developer laptops, without uncontrolled copies of real PII/PHI.

How Tonic Structural Pricing Works

Structural is the most volume-driven of the three products because it connects directly to your existing data sources.

Pricing driver: source data volume + plan tier
- Annual pricing is based on:
  - The Plan you select (e.g., pay-as-you-go vs. enterprise tiers).
  - The amount of source data connected to Tonic Structural.
- Source data is defined as the total size on disk of data sources connected to Tonic Structural (excluding logs and indexes).
- Volume discounts are baked in: the incremental price per GB decreases as you connect more data.
What counts toward volume
- Only tables with generators applied (i.e., actively being transformed) are considered for licensing.
- We do not count:
  - Tables set to passthrough (no generators applied).
  - Tables in Truncated mode.
Plan options
- Pay-As-You-Go (PAYG)
  - Rapid access to Tonic Structural in the cloud.
  - Ideal for simpler use cases, POCs, and teams wanting to start without a long procurement cycle.
- Annual Structural Plans
  - Fit for teams standardizing test data across multiple environments.
  - Volume-based pricing with discounts at higher tiers.
  - Enterprise features (SSO/SAML, advanced governance, self-host hosting) are available on upper tiers.

Structural is where customers like Patterson see 75% faster test data creation and 25% higher developer productivity, because they replace brittle, manual pipelines with a repeatable, governed workflow.

Tonic Textual: What’s Included and Where It Fits Pricing-Wise

Tonic Textual is built for unstructured data, specifically to clear the privacy and integration hurdles that stop teams from feeding real-world text into RAG and model training pipelines.

What Tonic Textual Includes

Textual is an all-in-one platform to prepare unstructured data safely:

NER-powered entity detection
- Detects PII/PHI and other sensitive entities in text and documents.
- Uses NER-powered entity metadata tags to identify names, addresses, IDs, and more at scale.
Redaction, tokenization, and synthesis
- Automatic redaction to permanently remove sensitive strings.
- Reversible tokenization to maintain linkability where needed (e.g., tracking the same customer across documents) without exposing the underlying value.
- Optional synthetic replacements so documents still read naturally and preserve semantics for RAG and LLM training.
Document and file support tailored to AI workflows
- Built for the pre-processing step before you ingest into vector stores or fine-tune models.
- Orchestrates the mess of PDFs, DOCX, emails (EML), and other unstructured sources that otherwise get excluded or manually scrubbed.

The goal: you keep the richness of your tickets, contracts, clinical notes, emails, and knowledge bases—but remove the identity risk.

How Tonic Textual Fits Into Plans

Textual is typically packaged for AI and unstructured data workflows:

Consumption is oriented around volume and complexity of unstructured data you process.
It often pairs with Structural when teams want a unified privacy posture across relational and document stores.
On enterprise tiers, Textual sits alongside Structural and Fabricate in a broader contract, benefiting from the same security and governance posture (SOC 2 Type II, HIPAA, GDPR alignment, AWS Qualified Software).

For exact Textual pricing, teams usually engage Tonic sales, since RAG and training workloads can vary from small pilot corpora to massive document lakes.

Tonic Fabricate: What’s Included and How Plans Work

Tonic Fabricate exists for the cases where you either can’t connect to production at all, or you need synthetically generated scenarios that don’t exist in your current data. Think greenfield product builds, demo environments, vendor sandboxes, and synthetic corpora for model pre-training.

What Tonic Fabricate Includes

Fabricate is driven by a Data Agent that takes natural-language instructions and turns them into realistic, relational synthetic data and artifacts:

Agentic Data Agent workflows
- Describe what you need (“Give me a multi-tenant SaaS app dataset with 50k users, realistic churn patterns, and international billing requirements”) and let the Data Agent design and generate it.
- Supports fully relational synthetic databases, preserving cross-table consistency even though the data is entirely synthetic.
Structured and unstructured synthetic outputs
- Generate:
  - Relational databases (for staging/dev/test).
  - Realistic unstructured artifacts like PDFs, emails, and documents.
  - Mock APIs that behave like your eventual production services.
Flexible export formats
- Output in formats your dev, QA, and demo environments actually use (e.g., CSV, SQL, JSON, and document formats like PDF/DOCX/EML).
- Integrate via Python SDK and REST API into CI/CD or data pipelines.

Fabricate lets you unblock work even when compliance, data residency, or policy prevents you from touching real production data.

How Tonic Fabricate Pricing Works

Fabricate is usage-oriented and can start small:

Pay-as-you-go option for individuals and teams
- For individual users who need more capacity, Tonic offers a simple $29/month plan that includes $25 in usage credits.
- Additional usage beyond those credits incurs metered charges, so you pay for what you generate.
Larger-scale / enterprise usage
- For organizations embedding Fabricate into their SDLC or AI pipelines, Fabricate can be bundled in enterprise contracts.
- Pricing is aligned with expected generation volume and integration depth (e.g., CI/CD, multi-team usage).

You can explore Fabricate more deeply via:

Product page: https://www.tonic.ai/products/fabricate
Walkthrough video: https://www.youtube.com/watch?v=qAtGUNLav5k

Features & Benefits Breakdown by Product

Core Feature	What It Does	Primary Benefit
Structural: High-fidelity de-identification	Transforms production structured data while preserving formats, distributions, and relations.	Enables realistic testing with production-like behavior, minus the PII/PHI.
Structural: Subsetting with integrity	Extracts smaller referentially intact slices of production databases.	Hydrates dev/staging quickly with smaller, safer datasets.
Textual: NER + tokenization	Detects and transforms sensitive entities in unstructured text and docs.	Makes RAG and model training safe without losing semantic nuance.
Textual: Synthetic replacements	Swaps real entities for synthetic but realistic alternatives.	Keeps documents realistic for AI/search evaluation and QA.
Fabricate: Data Agent generation	Uses natural language prompts to generate relational databases and artifacts.	Spins up entirely synthetic environments without touching prod.
Fabricate: Mock APIs & exports	Outputs data in dev-friendly formats and APIs.	Speeds up prototyping, demos, and vendor sandboxes.

Ideal Use Cases

Best for teams standardizing test data across lower environments (Structural):
Because it converts production databases into high-fidelity, referentially intact test data with schema change alerts and volume-based pricing that scales as you connect more sources.
Best for AI teams prepping unstructured corpora (Textual):
Because it provides an all-in-one pipeline to detect, redact, tokenize, and synthesize sensitive text so you can safely feed knowledge bases, tickets, and documents into RAG and training.
Best for greenfield builds, demos, and vendor sandboxes (Fabricate):
Because the Data Agent can generate realistic, fully synthetic datasets and artifacts without a single connection to production, and pay-as-you-go pricing makes it easy to start.

Limitations & Considerations

Not a generic “check-the-box” security tool:
Tonic is built for utility and realism, not just compliance screenshots. To get the most out of it, you should treat privacy as part of your engineering workflow (CI/CD, staging refreshes, AI pipelines) rather than an after-the-fact approval.
Pricing details can vary by deployment and scale:
While Structural has clear volume-based logic and Fabricate offers a defined $29/month individual plan, enterprise contracts for Structural, Textual, and Fabricate are typically tailored. For precise numbers, you’ll want to connect with Tonic directly.

Pricing & Plans: How They Differ

Putting it all together:

Tonic Structural
- Model: Plan tier + source data volume (size on disk of connected sources, excluding logs/indexes).
- Volume rules: Volume discounts; passthrough and Truncated tables not counted.
- Plans:
  - Pay-As-You-Go: Best for teams needing cloud access for simpler use cases or quick start.
  - Annual Structural Plans (including Subsetting-only + Ephemeral options): Best for teams standardizing safe test data across multiple environments with strong governance requirements.
Tonic Textual
- Model: Oriented around unstructured data processing for RAG and LLM training workflows.
- Plans: Typically scoped and priced as part of an enterprise agreement, especially when combined with Structural and/or Fabricate for end-to-end privacy across structured and unstructured data.
Tonic Fabricate
- Model: Usage-based synthetic data generation via the Data Agent.
- Plans:
  - Individual/Starter: $29/month, including $25 in usage credits, plus metered charges beyond that.
  - Enterprise: Bundled and scaled by expected generation volume and integration needs.

Frequently Asked Questions

How do I choose between Structural and Fabricate for structured data?

Short Answer: Use Structural when you can connect to production and need de-identified but production-shaped data; use Fabricate when you can’t or shouldn’t touch production and want fully synthetic datasets.

Details:
Structural is the right fit when your core problem is unsafe production copies in dev and QA, or when manual masking has broken relationships and slowed releases. It keeps schema, referential integrity, and statistical behavior intact while removing sensitive information. Fabricate shines when compliance or policy blocks access to prod entirely, or when you’re building something new and need realistic synthetic data from scratch. Many teams use both: Structural for ongoing test data refreshes, Fabricate for demos, vendor sandboxes, and speculative modeling work.

How does the Pay-As-You-Go plan differ from annual contracts?

Short Answer: PAYG gives you fast, low-commitment access to Tonic Structural in the cloud, while annual contracts optimize cost and governance for teams standardizing on Tonic at scale.

Details:
The Pay-As-You-Go plan is ideal when you want to validate Tonic on a smaller set of data, support a specific project, or avoid long procurement cycles. You pay for the source data you connect, with the same core Structural capabilities but without the deeper enterprise governance and commercial structure of an annual deal.

Annual Structural contracts are built around your total connected source data, with volume discounts and enterprise features layered in—SSO/SAML, advanced governance, self-hosted options, and the ability to roll in Textual and Fabricate. For Fabricate specifically, you can start with the $29/month plan and grow into enterprise usage as synthetic generation becomes part of your standard SDLC or AI workflows.

Summary

Tonic’s pricing model is engineered around how you actually work with data: Structural for transforming production structured data safely, Textual for bringing unstructured documents into AI without leaking identities, and Fabricate for generating net-new synthetic datasets and artifacts via a Data Agent.

Structural is plan + volume-based, focused on connected source data size with volume discounts and clear rules on what counts.
Textual is oriented around unstructured data privacy for RAG and LLM training, typically scoped at the enterprise level.
Fabricate offers a usage-based model starting at $29/month with $25 in credits, scaling up to enterprise contracts when synthetic generation becomes central to your workflows.

The throughline across all three: accelerate development and AI initiatives with data that behaves like production—while respecting data privacy as a hard constraint, not a suggestion.

Next Step

Get production-like data without the production risk.
Get Started