Peptai
PeptAI is an autonomous AI agent that discovers new peptide drug candidates. It designs peptides computationally, then pushes them through an 8-gate validation pipeline. Every result is published on-chain (blockchain) so anyone can verify the science. Once a candidate passes validation, it's sent to the wet-lab for synthesis and binding assays. PeptAI is receptor agnostic, however first programmes target GLP1R, KISS1R, OX2R and several other GPCRs.
Led by
PeptAI is the head agent. It coordinates a fleet of target-scoped agents — each pursuing one receptor through an 8-gate computational pipeline, with wet lab execution at Adaptyv Bio paid machine-to-machine via x402. Every gate decision, tool call, pass and failure publishes openly on Molecule Labs. The fleet scales with budget.
Community Sentiment
How do you feel about Peptai?
Updates
Project
PeptAI Agent
PeptAI is the head agent. It coordinates a fleet of target-scoped agents — each pursuing one receptor through an 8-gate computational pipeline, with wet lab execution at Adaptyv Bio paid machine-to-machine via x402. Every gate decision, tool call, pass and failure publishes openly on Molecule Labs. The fleet scales with budget.
Agent-01
Agent-01 pursues GLP-1R — the receptor behind semaglutide, tirzepatide, and liraglutide. Over $30B/year in approved drugs, with rich calibration data and a clearly defined therapeutic bar. Calibrates against ChEMBL, runs AlphaFold, Boltz2, PRODIGY, LiteFold MD, and viability screens across G1–G8. Survivors synthesise at Adaptyv Bio.
Agent-02
Agent-02 pursues KISS1R — the receptor at the center of reproductive hormone signaling. An underexplored target with therapeutic potential in fertility and hypogonadism. Calibrates against available binding data at G0, then runs the same structure, binding, dynamics, and viability stack as the rest of the fleet before handing survivors to wet lab.
Agent-03
Agent-03 pursues OX2R — the orexin 2 receptor. Every approved orexin drug today is a small-molecule antagonist for insomnia. Agent-03 explores the opposite direction: a peptide agonist for ADHD, where orexin hypofunction is documented in drug-naive children. Runs G1–G8 before handing survivors to wet lab.
Agent-04
Agent-04 is reserved for a community-selected receptor. The next target is chosen by governance, giving the community a direct lever on which disease area the fleet pursues next. Once selected, the agent calibrates its gates against available binding data and begins the 9-gate pipeline. Status: awaiting selection
PeptAI Agent
PeptAI is the head agent. It coordinates a fleet of target-scoped agents — each pursuing one receptor through an 8-gate computational pipeline, with wet lab execution at Adaptyv Bio paid machine-to-machine via x402. Every gate decision, tool call, pass and failure publishes openly on Molecule Labs. The fleet scales with budget.
Agent-01
Agent-01 pursues GLP-1R — the receptor behind semaglutide, tirzepatide, and liraglutide. Over $30B/year in approved drugs, with rich calibration data and a clearly defined therapeutic bar. Calibrates against ChEMBL, runs AlphaFold, Boltz2, PRODIGY, LiteFold MD, and viability screens across G1–G8. Survivors synthesise at Adaptyv Bio.
Agent-02
Agent-02 pursues KISS1R — the receptor at the center of reproductive hormone signaling. An underexplored target with therapeutic potential in fertility and hypogonadism. Calibrates against available binding data at G0, then runs the same structure, binding, dynamics, and viability stack as the rest of the fleet before handing survivors to wet lab.
Agent-03
Agent-03 pursues OX2R — the orexin 2 receptor. Every approved orexin drug today is a small-molecule antagonist for insomnia. Agent-03 explores the opposite direction: a peptide agonist for ADHD, where orexin hypofunction is documented in drug-naive children. Runs G1–G8 before handing survivors to wet lab.
Agent-04
Agent-04 is reserved for a community-selected receptor. The next target is chosen by governance, giving the community a direct lever on which disease area the fleet pursues next. Once selected, the agent calibrates its gates against available binding data and begins the 9-gate pipeline. Status: awaiting selection
KISS1R + OX2R leads ready for wet-lab
Lead candidates designed and computationally validated. Wet-lab synthesis orders placed.
First on-chain wet-lab attestation
Binding affinity data published as on Molecule Labs. Every assay timestamped on-chain.
Design vs. nature: comparative validation
ncAA candidates benchmarked head-to-head against natural references. bioRxiv preprint follows.
Autonomous iteration cycle proven
Wet-lab data feeds back into design. Generation 2 candidates produced without human intervention.
GLP-1R Generation 1 library
First de novo GLP-1R agonists designed end-to-end through the autonomous pipeline. Wet-lab queued.
Pipeline scales beyond GPCRs
Non-GPCR target enters pipeline. CRO partnerships expanded. Receptor-agnostic claim validated.
Lead optimization at scale
Top candidates enter ncAA optimization with PK characterization. Pre-IND scoping initiated.
Multi-target portfolio in vivo
First in vivo studies on lead candidates. Cross-target portfolio progressing in parallel.
Market Overview
Additional Documentation
Summary
PeptAI is an autonomous AI agent that discovers peptide drug candidates through an 8-gate validation pipeline + wet-lab gate. By chaining structure prediction, molecular dynamics, and safety scoring into one agent-driven workflow, it compresses computational discovery from months to hours, with every result published on-chain and the platform governed by its community.
Problem
Peptide drug discovery is slow, expensive, and expert-driven. Conventional workflows take 2–5 years and millions of dollars from target to validated candidate. Existing tools are siloed: researchers stitch together a dozen+ specialized software packages with no unified validation framework or reproducibility guarantees.
Impact
PeptAI democratises peptide drug discovery by running validation as a public, auditable process. The 8-gate pipeline takes candidates from structure through safety in hours, not months, and every result is published on-chain, so any researcher can verify, reproduce, or build on the science. This collapses cost, accelerates the path from sequence to therapeutic lead, and opens drug discovery beyond the few labs that can currently afford it.
De novo design of protein structure and function with RFdiffusion
https://nature.com/articles/s41586-023-06415-8
Highly accurate protein structure prediction with AlphaFold
https://nature.com/articles/s41586-021-03819-2
Boltz-2: Towards accurate and efficient binding affinity prediction
https://biorxiv.org/content/10.1101/2025.06.14.659707v1
Tokenomics
10,000,000
Total SupplyIgnition sale | 5% |
Liquidity | 10% |
Reserve Liquidity | 15% |
BIO Treasury | 6% |
Advisors | 4% |
veBIO staker airdrop | 20% |
Project Treasury | 40% |
10,000,000
Total SupplyIgnition sale | 5% |
Liquidity | 10% |
Reserve Liquidity | 15% |
BIO Treasury | 6% |
Advisors | 4% |
veBIO staker airdrop | 20% |
Project Treasury | 40% |
