/

Profluent

  • What it is:Profluent is a biotechnology company that uses frontier AI foundation models and wet-lab integration to design novel, functional proteins for biomedicine, therapeutics, agriculture, and industrial applications.
  • Best for:Biotech/pharma R&D teams, Academic researchers, Therapeutics developers
  • Pricing:Free tier available, paid plans from Custom licensing
  • Rating:85/100Very Good
  • Expert's conclusion:For those biotech innovators who are ready to utilize the frontiers of AI to create breakthrough protein designs for therapeutics and agriculture, Profluent is the perfect choice.
Visit website
Reviewed byMaxim Manylov·Web3 Engineer & Serial Founder

What Is Profluent and What Does It Do?

Profluent is a biotech firm that uses AI to develop new proteins for medicine, food production, and other industries, and designs them through a process called "protein design" and validates their function using computational models and molecular simulation. The company has the potential to fundamentally change how we use biology and apply it to our needs and desires by applying machine learning and biology to generate novel functional proteins.

Active
📍Emeryville, CA
📅Founded 2022
🏢Private
TARGET SEGMENTS
Biotech CompaniesPharmaceuticalsAgricultureBiomanufacturing

What Are Profluent's Key Business Metrics?

📊
$44M
Total Funding Raised
📊
$35M Series A
Latest Funding Round
📊
80 billion+
Protein Sequences in Database
📊
3.4 billion protein sequences
ProGen3 Training Data
📊
20 drug targets
OpenAntibodies Targets
📊
$660B historical sales
Market Addressed by OpenAntibodies

How Credible and Trustworthy Is Profluent?

85/100
Excellent

Profluent demonstrates a high degree of technical leadership in the field of AI-based protein design with validated results (OpenCRISPR-1, OpenAntibodies) along with significant investment from top venture capital firms and a demonstrated ability to scale biological AI models.

BREAKDOWN
Product Maturity82/100
Company Stability88/100
Security & Compliance70/100
User Reviews75/100
Transparency85/100
Support Quality80/100
TRUST SIGNALS
OpenCRISPR-1: AI-designed open-source gene editorOpenAntibodies: 20 novel antibody designs matching commercial therapiesBacked by Spark Capital, Insight Partners, Air Street CapitalProGen3: Billion-parameter models with wet-lab validationNamed Challenger in AI protein engineering platforms

What is the history of Profluent and its key milestones?

2022

Company Founded

Founded by Ali Madani and a group of experts in both AI and biology who developed the first AI-only protein design firm.

2022

Early Funding

Received its first round of funding and initiated the development of AI-driven deep generative models for protein design.

2024

$35M Series A

Received $35M in Series A funding from Spark Capital and Insight Partners, which brings the company's total funding to date to $44M.

2024

OpenCRISPR-1 Launch

Developed OpenCRISPR-1, the world's first AI-designed open source gene editor.

2025

ProGen3 Announcement

Launched the ProGen3 family of AI-based protein generation models that can be trained on 3.4 billion protein sequences, each having a billion parameters.

2025

OpenAntibodies Initiative

Designed novel antibodies for 20 disease targets for 7 million patients, with a combined market size of $660 billion.

What Are the Key Features of Profluent?

✨
ProGen3 Protein Generation
Can train trillion parameter language models on 3.4 billion protein sequences and design novel functional proteins in a single shot.
✨
OpenAntibodies Design
Can produce computationally-predicted antibodies that have similar or superior efficacy than commercially available therapies against 20 different disease targets, without encountering any patent issues.
✨
Scaling Laws Validation
Demonstrates for the first time the scalability of AI in biological design; the larger the model, the better the prediction accuracy.
✨
OpenCRISPR-1 Gene Editor
Releases ultra compact AI-designed open source CRISPR gene editor for academic researchers.
✨
80 Billion Protein Database
Has proprietary databases that far exceed what is available publicly and serves as the base for training the companies' foundation models.
✨
Partner Access Programs
Offers licensing agreements for the designed molecules, early access to the companies' foundation model API, and collaborative partnerships with the company.
✨
Natural Language Protein Specification
Future vision of being able to allow scientists to define the properties they desire of a protein using natural language, and then use an Artificial Intelligence system to create DNA code that would produce such a protein.

What Technology Stack and Infrastructure Does Profluent Use?

Infrastructure

Cloud-based AI training infrastructure

Technologies

Deep Generative ModelsLarge Language ModelsProtein Language Models

Integrations

Therapeutics PartnersAgriculture PartnersBiomanufacturing Partners

AI/ML Capabilities

ProGen3 family of billion-parameter protein language models trained on 3.4B+ protein sequences from 80B sequence database, demonstrating scaling laws with wet-lab validated single-shot protein generation including antibodies and gene editors

Based on company website, technical announcements, and industry reports

What Are the Best Use Cases for Profluent?

Biotech Drug Discovery Teams
Create new antibodies that match commercially available antibodies for +20 different targets in a single shot process and speed up the therapeutic development cycle by avoiding patent issues associated with OpenAntibodies.
Gene Therapy Researchers
Gain early access to OpenCRISPR-1 ultra-compact AI designed gene editors and foundational models to design your own customized gene editors through early access programs.
Pharmaceutical R&D Departments
License pre-designed proteins or develop bespoke proteins for your company based on specific therapeutic targets you want to address with our validated AI design platform.
Agricultural Biotech Companies
Design novel proteins for crop protection, yield enhancement and sustainability applications using our AI powered protein design platform.
Biomanufacturing Firms
Use our AI to design and engineer custom enzymes and proteins for industrial processing, and/or use our proteins to enable the production of sustainable materials and manufacturing applications.
NOT FORReal-time Diagnostic Companies
Not suitable – protein design timelines do not meet real time diagnostic requirements that require a turnaround of hours or less.
NOT FORSmall Academic Labs
Limited access – we have enterprise focused licensing and partnership programs as opposed to tools for individual researchers.

How Much Does Profluent Cost and What Plans Are Available?

Pricing information with service tiers, costs, and details
☐Service$CostℹDetails🔗Source
OpenCRISPR-1$0Freely available open-source gene editor for researchers and commercial useProfluent website
OpenAntibodiesRoyalty-free or single upfront licensing fee20 antibodies for drug targets addressing 7 million patients, DNA recipes publicly availableProfluent ProGen3 showcase
Model LicensingCustom licensingEarly access program for ProGen3 models and other AI protein generation modelsProfluent website
CollaborationsCustom quotePartnerships for custom protein design in therapeutics, agriculture, biomanufacturing—
OpenCRISPR-1$0
Freely available open-source gene editor for researchers and commercial use
Profluent website
OpenAntibodiesRoyalty-free or single upfront licensing fee
20 antibodies for drug targets addressing 7 million patients, DNA recipes publicly available
Profluent ProGen3 showcase
Model LicensingCustom licensing
Early access program for ProGen3 models and other AI protein generation models
Profluent website
CollaborationsCustom quote
Partnerships for custom protein design in therapeutics, agriculture, biomanufacturing

How Does Profluent Compare to Competitors?

FeatureProfluentAlphaFoldBIOVIA Discovery StudioIsomorphic Labs
Core FunctionalityAI protein generation & designProtein structure predictionMolecular modeling & simulationDrug discovery AI
PricingCustom licensing/partnershipsFree (open source)Commercial licenseProprietary (Alphabet)
Free Tier AvailabilityYes (OpenCRISPR, OpenAntibodies)YesNoNo
Enterprise FeaturesPartnerships with pharma—YesYes
API AvailabilityEarly access modelsNoYesProprietary
Data Scale3.4B+ protein sequences214M structuresResearch databasesDeepMind data
Support OptionsPartnership supportCommunityVendor supportEnterprise
Open Source ComponentsYes (OpenCRISPR-1)YesNoNo
Therapeutic ApplicationsAntibodies, gene editingStructure predictionDrug optimizationDrug discovery
Scaling Laws DemonstratedYes (ProGen3)NoNoPartial
Core Functionality
ProfluentAI protein generation & design
AlphaFoldProtein structure prediction
BIOVIA Discovery StudioMolecular modeling & simulation
Isomorphic LabsDrug discovery AI
Pricing
ProfluentCustom licensing/partnerships
AlphaFoldFree (open source)
BIOVIA Discovery StudioCommercial license
Isomorphic LabsProprietary (Alphabet)
Free Tier Availability
ProfluentYes (OpenCRISPR, OpenAntibodies)
AlphaFoldYes
BIOVIA Discovery StudioNo
Isomorphic LabsNo
Enterprise Features
ProfluentPartnerships with pharma
AlphaFold—
BIOVIA Discovery StudioYes
Isomorphic LabsYes
API Availability
ProfluentEarly access models
AlphaFoldNo
BIOVIA Discovery StudioYes
Isomorphic LabsProprietary
Data Scale
Profluent3.4B+ protein sequences
AlphaFold214M structures
BIOVIA Discovery StudioResearch databases
Isomorphic LabsDeepMind data
Support Options
ProfluentPartnership support
AlphaFoldCommunity
BIOVIA Discovery StudioVendor support
Isomorphic LabsEnterprise
Open Source Components
ProfluentYes (OpenCRISPR-1)
AlphaFoldYes
BIOVIA Discovery StudioNo
Isomorphic LabsNo
Therapeutic Applications
ProfluentAntibodies, gene editing
AlphaFoldStructure prediction
BIOVIA Discovery StudioDrug optimization
Isomorphic LabsDrug discovery
Scaling Laws Demonstrated
ProfluentYes (ProGen3)
AlphaFoldNo
BIOVIA Discovery StudioNo
Isomorphic LabsPartial

How Does Profluent Compare to Competitors?

vs AlphaFold (DeepMind)

The focus of Profluent is generative protein design whereas AlphaFold is excellent at predicting the structure of proteins. In addition, Profluent has used 3.4B protein sequences versus AlphaFold’s 214M structures to enable the creation of completely novel proteins beyond those predicted.

Profluent for de novo protein design; AlphaFold for structure analysis.

vs Isomorphic Labs

Both companies are applying Large Language Model (LLM) technology to Biology but Profluent emphasizes open source outputs (OpenCRISPR) and has shown scaling laws in protein design whereas Isomorphic is benefiting from the vast resources of Alphabet but is generating much less publicly accessible data.

Profluent provides new drugs with novel protein structures; Isomorphic for fully-integrated drug discovery platforms.

vs BIOVIA Discovery Studio

Molecular modeling as it traditionally exists versus Profluent’s AI native generative approach to protein design. BIOVIA is more suited to traditional established simulation workflows, where as Profluent is more capable of producing novel, patent free proteins.

BIOVIA for physics-based optimization; Profluent for AI generated sequences.

vs EvolutionaryScale

Both companies were started by ex-hyperscaler AI researchers working on protein LMs. Profluent has significantly greater traction in terms of open-source momentum (OpenCRISPR has been adopted worldwide), and a significantly larger protein database (115B via Protein Atlas).

Profluent leads in open innovation; EvolutionaryScale competitive in closed models.

What are the strengths and limitations of Profluent?

Pros

  • The group has demonstrated scaling laws — larger models can be used to predictably improve protein performance.
  • A massive protein database — 115 billion unique proteins — the largest database in the world.
  • Open source leadership — OpenCRISPR-1 is being used widely by the pharmaceutical and research communities.
  • Novel, patent-free proteins — OpenAntibodies are being designed to avoid patent disputes.
  • Proven functional — the company's AI generated proteins are showing to match or exceed commercial antibodies in terms of function.
  • Funding momentum — $150 million in funding overall, Bezos Expeditions investing.
  • Wide range of applications — therapeutics, agriculture, biomanufacturing.

Cons

  • Commercialization stage — currently in the research phase with most partnerships/partnership style collaborations — no therapeutics have been approved yet.
  • Platform based around research — users need to have biology knowledge to use the platform.
  • Custom pricing — there are no standard SaaS pricing tiers available to customers.
  • No public API — only providing early access to their models.
  • User reviews — limited — this is a biotech startup — very few independent evaluations have been made available.
  • Experimental validation needed — while some wet lab results have shown promise — it is still early days for the technology.
  • Partnership dependent — Profluent needs partners to deliver applications.

Who Is Profluent Best For?

Best For

  • Biotech/pharma R&D teams — Provides access to novel, patent-free antibodies and gene editors which will accelerate discovery.
  • Academic researchers — OpenCRISPR-1 and future OpenAntibodies will provide researchers with the opportunity to perform cutting edge experimentation at no cost.
  • Therapeutics developers — ProGen3 designs antibodies similar to the $660 billion dollar market of therapeutics.
  • Agriculture/biomanufacturing companies — Provides custom enzymes/proteins for industrial applications.
  • AI/biology hybrid teams — Has proven scaling laws — a massive 115 billion protein database.

Not Suitable For

  • Non-expert biologists — Requires advanced knowledge of protein engineering — consider using pre-existing protein engineering tools if you do not have this knowledge.
  • Small startups seeking SaaS — Uses a custom partnership model — does not offer a self-serve platform — look for a well-established platform instead.
  • Structure prediction only — Can use free version of AlphaFold — Profluent focuses on generating/designing proteins.
  • Budget-constrained researchers — Despite many parts being open source, commercial licenses are still expensive.

Are There Usage Limits or Geographic Restrictions for Profluent?

Model Access
Early access program and partnerships only
Protein Database
115 billion unique proteins (Protein Atlas)
Open Source Assets
OpenCRISPR-1 freely available, OpenAntibodies royalty-free/upfront licensing
Training Data Scale
ProGen3 trained on 3.4B sequences
Commercial Availability
Partnerships required for production applications
Geographic Availability
US-based, global research partnerships

Is Profluent Secure and Compliant?

Data SecurityProprietary protein sequence databases with research/commercial partnerships
Open Source LicensingOpenCRISPR-1 free for research and commercial use
Enterprise PartnershipsWorks with major pharma, adopted by thousands of commercial operations
IP ProtectionGenerates patent-free novel proteins different from natural/patented sequences
Research ComplianceAcademic and pharma collaborations with established compliance frameworks

What Customer Support Options Does Profluent Offer?

Channels
Partnership inquiriesCustom project partnershipsOpen source researcher supportModel access applications
Hours
Business hours
Response Time
Partnership-dependent
Satisfaction
N/A (research platform)
Specialized
Dedicated partnership teams for biotech/pharma clients
Business Tier
Custom support for commercial collaborations
Support Limitations
•No self-serve customer support portal
•Enterprise/pharma partnership model only
•No public ticketing system

What APIs and Integrations Does Profluent Support?

API Type
Early access API available for foundation models (ProGen3 family). Likely REST-based for protein generation and customization.
Authentication
Not publicly detailed. Enterprise-level authentication expected for early access program and model licensing.
Webhooks
No public information on webhook support.
SDKs
No official SDKs mentioned. OpenCRISPR-1 available as open-source for integration.
Documentation
Limited public documentation. Early access program provides API access for select partners (profluent.bio/showcase/progen3).
Sandbox
No public sandbox. Early access requires partnership application.
SLA
Not publicly disclosed. Enterprise partnerships likely include custom SLAs.
Rate Limits
Not publicly available. Enterprise-specific limits expected.
Use Cases
Protein generation, antibody design (OpenAntibodies), custom model fine-tuning for therapeutics, agriculture, and biomanufacturing applications.

What Are Common Questions About Profluent?

Profluent utilizes Frontier AI foundation models that have been trained on 3.4 + billion protein sequences to generate novel proteins de novo. ProGen3 models allow for a one shot design of antibodies as well as custom gene editors such as OpenCRISPR-1.

The world's first open-source, AI generated gene editor was developed and made publicly available by Profluent; OpenCRISPR-1 has greater than 400 mutations compared to natural Cas9. OpenCRISPR-1 may be utilized for both commercial and ethical research purposes.

Partners may license pre-designed molecules, develop co-branded protein solutions or utilize the early access API program for the foundation models. Access to these programs will require an application via profluent.bio.

Profluent develops proteins for therapeutics (antibodies, gene editors), agriculture and biomanufacturing. Most recently Profluent has focused its efforts towards OpenAntibodies for 20 drug targets as well as partnerships for Rett Syndrome genomic medicines.

Profluent utilizes large language models to discover new regions of the protein sequence space that are unexplored by nature. Additionally, they demonstrate that biological design may follow scaling laws, optimizing multiple characteristics at once.

Yes, Profluent has published in peer-reviewed journals including Nature Biotechnology, the first AI designed CRISPR system and thousands of researchers and pharmaceutical companies have utilized OpenCRISPR-1.

Profluent's Protein Atlas contains over 115 billion unique proteins which makes it claimed to be the worlds' largest protein database. This database is utilized to train Profluent's AI foundation models.

Profluent currently partners with the following organizations; Revvity, Corteva Agrisciences, Rett Syndrome Research Trust, Integrated DNA Technologies and Ensoma. These partnerships are primarily in the fields of therapeutics, agriculture and gene editing.

Is Profluent Worth It?

Profluent has pushed the limits of artificial intelligence in protein design by making groundbreaking advancements such as the world’s first artificially generated CRISPR gene editing system and the world's first single-shot antibody design for major drug targets. The funding of their $106 million Series B round and the creation of the Protein Atlas that includes 115 billion proteins will make Profluent a leading force in programmable biology. Furthermore, early access APIs and open source software development represent an ongoing commitment to building an ecosystem around Profluent.

Recommended For

  • Drug development companies which are creating new drugs using innovative methods require customized antibodies or gene editors.
  • Companies engaged in agribiotechnology that need to develop the most efficient enzymes and proteins for agriculture.
  • Pharmaceutical Research and Development (R&D) teams looking to use Artificial Intelligence (AI)-first approaches to design proteins at large scales.
  • Researchers who want to advance the study of computational biology.

!
Use With Caution

  • Production ready commercial proteins right away – currently almost all of these are still R&D platforms.
  • Organizationally, they have no experience with AI or biotech -- they will need a technical partner.
  • They have budget constraints – Profluent uses an enterprise pricing model.

Not Recommended For

  • Small molecule pharmaceutical research and development teams – they are only focused on proteins.
  • Teams which need off the shelf protein reagents – they are focused on designing proteins.
  • Technical buyers who do not have a background in science – they will require a scientific partnership.
Expert's Conclusion

For those biotech innovators who are ready to utilize the frontiers of AI to create breakthrough protein designs for therapeutics and agriculture, Profluent is the perfect choice.

Best For
Drug development companies which are creating new drugs using innovative methods require customized antibodies or gene editors.Companies engaged in agribiotechnology that need to develop the most efficient enzymes and proteins for agriculture.Pharmaceutical Research and Development (R&D) teams looking to use Artificial Intelligence (AI)-first approaches to design proteins at large scales.

What do expert reviews and research say about Profluent?

Key Findings

Profluent is pioneering AI-based protein design utilizing ProGen3 models that were trained on 3.4 billion sequence data sets, the first AI-designed gene editor, OpenCRISPR-1, and OpenAntibodies for 20 drug targets. Profluent raised a $106 million Series B round in December 2025 from Altimeter and Bezos Expeditions to grow its presence in therapeutics, agriculture, and biomanufacturing. Profluent’s partnerships with Revvity, Corteva, and RSRT confirm that the company has achieved commercial success; Profluent’s Protein Atlas contains 115 billion proteins.

Data Quality

Good - comprehensive information from official website, recent funding press releases, and peer-reviewed publications. Limited details on commercial pricing/API terms (enterprise sales contact required).

Risk Factors

!
Profluent is a relatively young early-stage company formed in 2022 and its technology is still growing.
!
There are many other competitors that are already established in the highly competitive area of protein engineering.
!
Profluent faces regulatory challenges when it comes to developing AI-based therapeutics.
!
Profluent depends upon the continued validity of the fundamental principles of how AI grows. I will make the following text sound more like a human speaking. I will not answer questions or add any information; I can only paraphrase the text you provide to me as follows: Begin Text
Last updated: February 2026

What Additional Information Is Available for Profluent?

Recent Funding

The $106M Series B investment round of December 2025 is led by Altimeter Capital and Bezos Expeditions. Overall, this has brought total funding to more than $200 million from prior backers Spark Capital and Insight Partners. This funding will be used to expand the platform in therapeutics, agriculture and biomanufacturing.

Scientific Achievements

The first time that Large Language Models (LLMs) generated functional proteins was demonstrated (in Nature Biotechnology 2023). The group also created the first AI-designed CRISPR and demonstrated that the scaling laws that are known to exist for protein design do indeed exist. The Protein Atlas contains approximately 115 billion unique proteins which makes it the world’s largest.

Open Source Initiative

The OpenCRISPR-1 is free of charge for all researchers and commercial uses. It is being utilized by thousands of researchers worldwide (including major pharmaceutical companies) and includes synthetic guide RNA. This demonstrates that the AI-designed editors have performance capabilities equal to those of the natural Cas9.

Key Partnerships

Therapeutics: Rett Syndrome Research Trust (custom base editors) and Ensoma; Agriculture: Corteva Agriscience; Commercial: Revvity and Integrated DNA Technologies. There is an early access API program available.

Founder & Leadership

The company was founded in 2022 by Ali Madani (Chief Executive Officer). The team consists of experts in both AI/ML and biology and includes Peter Cameron (Senior Vice President, Gene Editing) who is responsible for the translation of their technology.

What Are the Best Alternatives to Profluent?

  • •
    Generate Biomedicines: The company has an AI-protein therapeutics platform using Generative AI for the creation of antibodies and enzymes. Their clinical pipeline appears to be more mature than that of Generate Biomedicines, however they appear to be less committed to an open source approach. They may best be suited for companies that want to reduce the risks associated with therapeutics development.
  • •
    Absci: The company has an AI designed antibody and protein using generative models. They are a publicly traded company with clinical assets and a more traditional biotech approach. As such, they may best be suited for investors looking for exposure to the AI biotech sector through the public markets.
  • •
    Crisp Therapeutics: The company has an AI-driven gene editor design platform with a strong focus on protein engineering. They are direct competitors in the area of compact editors, however their ability to scale their designs using foundation models is limited compared to the company InstaDeep. They may best be suited for companies focused on gene therapy.
  • •
    Instadeep (BioNeMo): The company has protein design models supported by NVIDIA (such as MegaMolBART) and enterprise API access. Their offerings seem to be more focused on providing support for the underlying infrastructure rather than directly supporting protein design. They may best be suited for groups that already utilize NVIDIA GPUs for other purposes.
  • •
    Isomorphic Labs: Use of an Alphabet-backed structure prediction and protein design model called AlphaFold3. Focuses on the design of proteins based on their 3-D structure, rather than the sequence-based design process used by Profluent’s LLMs. Most suitable for structure-based drug development. (https://www.isomorphiclabs.com/)

Design Success & Validation Metrics

Hundreds of mutations away from known natural proteins
Sequence Design Novelty
Comparable to SpCas9 baseline
On-Target Editing Efficiency
Higher relative to SpCas9
Off-Target Specificity
20 distinct drug targets designed
Antibody Design Coverage
7 million patients historically treated
Patient Population Addressed

Supported Generative Models

Protein Language Models

Training of a billion parameter transformer-based model on 3.4 billion protein sequences containing 1.1 trillion individual amino acids in each sequence.

ProGen3 Architecture

Design of sparse architectures that result in a fourfold increase in training speed without compromising modeling performance.

Full-Length Protein Generation

De novo creation of complete new proteins with no constraints placed upon their structure.

Domain-Specific Redesign

Targeted redesign of selected protein domains to enhance their functional properties while preserving the overall structure of the parent protein.

Multi-Property Optimization

Simultaneous fine-tuning of the binding affinity, stability, manufacturability, and immunogenicity of the redesigned protein.

Epitope-Matched Design

Creation of antibodies targeted against the same epitope as existing therapeutics, but having a completely different amino acid sequence.

Computational Evaluation Filters

Binding Affinity PredictionProtein Stability AssessmentManufacturability ScoringImmunogenicity PredictionSpecificity EvaluationSequence Diversity AnalysisFitness Landscape ModelingSolubility PredictionAggregation Risk AssessmentThermal Stability PredictionExpression Level OptimizationCodon OptimizationOff-Target Interaction ScreeningDomain Architecture ValidationEpitope Preservation Scoring

Supported Design Applications

Antibody Engineering
Design of therapeutic antibodies with optimized binding affinity, specificity, stability, and manufacturability
Gene Editing Systems
Custom CRISPR-Cas proteins, base editors, and prime editors optimized for precision, efficiency, and reduced off-target effects
Ultra-Compact Gene Editors
Miniaturized editing systems with reduced size constraints for improved delivery
Multi-Specific Therapeutics
Engineering of proteins with multiple binding specificities or functionalities
Difficult-to-Drug Targets
Solutions for cryptic or challenging protein targets through sequence space exploration
Therapeutic Development
Accelerated discovery and refinement of next-generation biologics for medicine and agriculture

Supported Data Modalities & Input Formats

Data ModalityInput FormatRequired/OptionalIntegration Depth
Protein SequencesFASTA, sequence databases, UniProt referencesRequiredPrimary training and design input from 3.4 billion protein sequences
Natural Protein ExamplesKnown antibodies, CRISPR systems, approved therapeuticsRequiredReference sequences for epitope matching and functional validation
Functional SpecificationsTarget specifications, binding requirements, functional constraintsOptionalGuides generation toward desired therapeutic properties
Manufacturing RequirementsProduction scalability specifications, storage stability needsOptionalEnables optimization for real-world therapeutic settings
Licensing ModelsAsset licensing, collaborative partnerships, early access requestsOptionalFlexibility in technology access and commercialization

Computational Requirements & Infrastructure

Model Scale
Billion-parameter language models with sparse architecture for efficiency
Training Data
3.4 billion full-length proteins and 1.1 trillion amino acid tokens (Profluent Protein Atlas v1)
Inference Capability
Single-shot design of complete antibodies and gene editors without iterative optimization
Deployment Options
Cloud-based platform, licensing models for partners, early access programs available
Platform Architecture
AI-first protein design platform optimized for speed and accuracy
Throughput
Capable of designing diverse proteins across multiple modalities and therapeutic targets

Regulatory Compliance & Scientific Validation

Nature Publication ValidationOpenCRISPR-1 design and characterization published in Nature with peer-reviewed methodology
Functional Validation in Human CellsAI-designed gene editors successfully edited human genome in HEK293T cells with high specificity
Performance BenchmarkingCompared against naturally occurring CRISPR systems and traditional protein engineering approaches
Immunogenicity AssessmentDesigned proteins show improved immunogenicity profiles relative to natural variants
Open Science CommitmentOpenCRISPR-1 and CRISPR-Cas Atlas released as open-source for academic and industry adoption
Broad Research AdoptionTens of thousands of academic and industry researchers accessing and validating designs
Intellectual Property FrameworkClear partnership models for licensing assets and collaborative development

Laboratory & Partnership Integration

Licensing Model

Available licensing of the designed proteins for use in various applications such as antibodies and gene editors.

Collaborative Partnerships

Pharmaceutical and biotechnology companies may engage directly with Isomorphic Labs for customized protein design.

Early Access Programs

Qualified partners will have structured access to Isomorphic Lab’s ProGen3 models and the foundational technology platform.

Open-Source Release

The open-source software platforms OpenCRISPR-1 and CRISPR-Cas Atlas are now publicly available to provide universal access to gene editing innovations.

One-Stop-Shop Model

Provides integrated solutions for gene editors, antibodies and other protein therapeutics tailored to meet the needs of partner companies.

Flexible Commercialization

Offers attractive economic and flexible partnering models for therapeutic, biomanufacturing, agricultural and climate-related applications.

Expert Reviews

📝

No reviews yet

Be the first to review Profluent!

Write a Review

Similar Products

EvolutionaryScale
evolutionaryscale.ai
Innate
innate.bio