Inferensys

Glossary

Soulbound Token

A Soulbound Token (SBT) is a non-transferable digital identity token representing the commitments, credentials, and affiliations of a person or entity, forming the basis for a decentralized, non-financialized reputation system.
Developer building agentic RAG system, retrieval pipeline diagram on laptop, technical workspace with notes.
NON-TRANSFERABLE IDENTITY

What is a Soulbound Token?

A Soulbound Token (SBT) is a non-transferable digital identity token representing the commitments, credentials, and affiliations of a person or entity, forming the basis for a decentralized, non-financialized reputation system.

A Soulbound Token is a publicly verifiable, non-transferable digital record issued to a blockchain address (a 'Soul') that represents an entity's provenance, affiliations, and achievements. Unlike fungible or non-fungible tokens (NFTs) that can be traded for financial value, an SBT is permanently bound to its recipient, functioning as an extended, decentralized resume or reputation certificate that cannot be sold or separated from the specific identity.

Proposed by Vitalik Buterin, Pooja Ohlhaver, and E. Glen Weyl in the 2022 whitepaper 'Decentralized Society: Finding Web3's Soul,' SBTs enable a composable Web of Trust where trust is transitive and non-financialized. By aggregating verifiable credentials—such as educational degrees, employment history, or medical records—into a single Reputation Graph, SBTs provide a robust mechanism for Sybil Resistance and Reputation Bootstrapping without relying on centralized identity providers or stake-weighted governance.

NON-TRANSFERABLE IDENTITY

Key Features of Soulbound Tokens

Soulbound Tokens (SBTs) form the cryptographic backbone of a decentralized, non-financialized reputation system. Unlike traditional assets, they are permanently bound to a specific account, representing commitments, credentials, and affiliations.

01

Permanent Non-Transferability

The defining characteristic of an SBT is its non-transferable nature. Once issued to a 'Soul' (a blockchain account), it cannot be moved or sold to another entity. This prevents the financialization of reputation and ensures that credentials and affiliations are authentic representations of a specific entity's history, not tradable assets. This is enforced at the smart contract level, disabling standard transfer functions.

02

Verifiable Credentialing

SBTs serve as a robust mechanism for verifiable credentials. An issuing authority (a university, DAO, or employer) can mint an SBT directly to a recipient's Soul. This token acts as a tamper-proof, cryptographically signed attestation of a fact, such as:

  • A university degree
  • A professional certification
  • Membership in a decentralized organization
  • Proof of attendance at an event
03

Sybil-Resistant Identity

By aggregating a collection of non-transferable tokens, a Soul builds a provable, unique identity history. This creates a powerful Sybil resistance mechanism. It becomes computationally and socially expensive for a single entity to fake a comprehensive, long-lived reputation across multiple pseudonymous accounts, as they cannot simply buy or transfer the necessary credentials to a new identity.

04

Decomposable Reputation Graph

SBTs enable the construction of a reputation graph that maps the relationships and trust between Souls and issuers. This graph is not a single score but a rich, decomposable network. A protocol can query this graph to assess a Soul's trustworthiness for a specific context, such as verifying a credit history without revealing the underlying transactions, or confirming DAO participation without exposing other affiliations.

05

Programmable Privacy with ZK-Proofs

While SBTs are public by default, advanced implementations integrate Zero-Knowledge Proofs (ZKPs). This allows a Soul to prove a claim about its SBTs without revealing the tokens themselves. For example, a Soul could prove 'I am a verified human with a credit score over 700' to a lending protocol without disclosing their exact score, wallet address, or the specific institutions that issued the underlying credentials.

06

Community-Governed Issuance and Revocation

The authority to issue and revoke SBTs is often governed by the community or a decentralized organization. Issuers can be other Souls or smart contracts. Crucially, revocation or a slashing condition can be implemented. If an entity acts maliciously, the issuing authority can burn the SBT representing their membership or certification, dynamically updating the reputation graph in response to real-world behavior.

SOULBOUND TOKENS EXPLAINED

Frequently Asked Questions

Clear, technical answers to the most common questions about non-transferable digital identity tokens and their role in decentralized reputation systems.

A Soulbound Token (SBT) is a publicly verifiable, non-transferable digital token representing the commitments, credentials, and affiliations of a specific Soul—a blockchain account or wallet. Unlike standard fungible or non-fungible tokens, SBTs are permanently bound to their recipient's address and cannot be sold, traded, or transferred to another entity. They function as an extended, decentralized resume, encoding facts such as educational degrees, employment history, professional certifications, and membership proofs directly on-chain. The mechanism relies on a Soul (the holder) and an Issuer (a trusted institution or entity) who cryptographically signs the token. Once issued, the SBT's non-transferability is enforced at the smart contract level, preventing any transferFrom or approve function from executing. This creates a persistent, tamper-evident record of an individual's provenance and social capital, forming the foundational layer for a Decentralized Society (DeSoc) where reputation is native, verifiable, and non-financialized.

Prasad Kumkar

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.