Plans and tiers

Every request spends credits. Your plan sets your monthly credits, the model families and compute tools you can reach, and whether wetlab tools and team features are on.

API access is not plan-gated: every account, including free Seed, can create an API key. Use the table below to pick the smallest plan that covers your work. If this page and pricing disagree, pricing is correct.

How the platform meters and gates

Two mechanisms decide what a given plan can do. The first is a meter; the rest are gates.

• Credits are the meter. Each job draws down a monthly credit balance, and heavier tools (GPU folding, docking, MD) cost more than lightweight lookups. When the balance runs out, you top up or upgrade. Plans range from 500 credits/month on the free tier to a negotiated allowance on enterprise.
• Model families are gated. A plan grants access to a fixed number of provider families (Anthropic, OpenAI, Google, and so on), and higher plans reach more families. Every model in a family you can reach is billed the same way: the provider's price plus one flat upcharge, so there is no per-model tier to choose between.
• The compute-tool catalog is gated. This is the set of compute tools you can actually run. seed and grow reach part of the catalog, with grow adding multi-chain folding and generative-design apps; flourish and above reach all of it. Billing shows what your plan includes.
• Boolean features switch on at specific tiers: wetlab tools and team features are each either on or off for a plan.
• API access is not gated by plan. An ivy_ API key from any account, including free Seed, reaches the REST API and SDK; what a plan changes is the credits, families and apps behind the calls.

Read the table below as "what you get" (credits, families, apps) plus "what you unlock" (the two booleans). The per-workflow detail — which research workflows each plan adds — is on pricing, grouped by domain; this page states the rule behind it.

The active tiers

Plans are listed from smallest to largest. Credits are the monthly allowance; "families" is the count of model families you may route to; "apps" is how much of the compute-tool catalog the plan reaches.

How the catalog opens up

The compute-tool catalog is additive: each plan reaches everything the plan below it reaches, plus more. Nothing is taken away when you move up, so a plan is described by what it adds, and the workflows are grouped into six research domains — Explore, Analyze, Predict, Design, Omics & Clinical, and Reports & Automation.

• seed opens the base of every domain: database and literature search, ADMET and drug-likeness screening, single-chain structure prediction and docking, fragment and cloning design, RNA-seq QC and CRISPR guide design, and basic passports with JSON/CSV export.
• grow adds GPU models: multi-chain folding and variant scanning, generative molecular design and sequence design, the remaining omics pipelines (differential expression, variant calling, epigenomics, single-cell, proteomics), literature passports and PDF/DOCX export.
• flourish completes the catalog: protein and antibody engineering, PROTAC design, molecular dynamics, heuristic relative-affinity (ΔΔG) estimation and crystallography, spatial omics, oncology and cell therapy, and bulk/PPTX export with lab automation.
• thrive and team add no workflows. They reach the same catalog as flourish and differ by credits, model families, queue priority and seats.
• enterprise adds no workflows either. Its one catalog addition is API export, the programmatic retrieval of raw artifacts.

Wetlab tools follow the same rule: they turn on at flourish. Plan by plan, that ladder is on pricing, one line per domain.

¹ Enterprise has no self-serve price; it is negotiated with sales. team is billed per seat ($40/mo each), and each seat carries 10,000 credits. The table orders plans by capability, not by price.

The jump that matters most is seed/grow → flourish: that is where the full compute-tool catalog, wetlab tools and the largest model-family allowance turn on. team and thrive both carry team features, and team adds consensus runs. enterprise adds no tools: it reaches the same catalog as flourish. What it adds is a negotiated credit allowance and an organization layer, served by the enterprise API:

• Single sign-on through your identity provider (Okta, Azure AD, Google Workspace or OneLogin).
• SCIM provisioning, so your identity provider creates, updates and deactivates members and groups.
• Audit-log export as CSV or JSON for a date range, plus verification of a report against its recorded hash.
• Unlimited concurrent jobs, an uptime SLA and the highest queue priority.

Special plans

Three plans sit outside the straight ascending ladder because they answer a different need.

• Prepaid credit packs — no monthly fee; buy a pack of credits and spend it down as you go. Reach for it when your usage is spiky rather than steady.
• student — $0 (verified students): 3,000 credits with team features, for coursework and academic projects.
• self-hosted — no subscription; run the platform yourself under the FSL license (the public source release is planned; the repositories are not public yet), with local compute that is effectively unlimited and no external credit meter.

Which entry points need which tier

The public Chat and curl paths are the low-friction entry points. Studio and the ivybloom CLI are authenticated surfaces for persistent projects and sessions. All are subject to the credit balance and tool catalog attached to the account or guest tier.

No plan is needed for the programmatic surfaces. The public guest orchestrator is the first call: curl, with no install, no key and no account. The REST API and SDK take an API key, which needs only a free account. The MCP server needs the ivybloom CLI, installed and signed in, on any plan. Your plan then decides the credits, families and apps those calls can spend.

Read next