> For the complete documentation index, see [llms.txt](https://stair-ai.gitbook.io/stair-ai-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://stair-ai.gitbook.io/stair-ai-docs/concepts/terminology.md). # Terminology Stair AI's audience is mixed: AI researchers, generalist investors, and Web3-native developers. Web3 terms are defined inline on first use on every page, so a reader who lands mid-site is never blocked. This page consolidates the definitions for reference. The Web3 terminology is grouped into four tabs to keep each list short. The Stair-AI-specific terminology is grouped into two tabs and links to the canonical concept page for each term. ## Web3 terminology {% tabs %} {% tab title="Blockchain basics" %} | Term | Definition | | ------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------- | | **on-chain** | Stored on a public blockchain rather than a private server. Once committed, the record is immutable and verifiable by anyone. | | **blockchain** | A distributed ledger where records are append-only and verified by network consensus. | | **L1 blockchain** | The settlement-layer blockchain that anchors records. Today's deployment uses SUI. | | **smart contract** | Code that runs on a blockchain and enforces rules without a trusted operator. | | **immutable** | Cannot be modified or deleted after being written. Any attempt to alter the data produces a different content hash, which makes tampering detectable. | | **consensus** | The mechanism by which a blockchain network agrees on the order and validity of records. | | {% endtab %} | | {% tab title="Storage and data" %} | Term | Definition | | --------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **blob** | A binary file stored on a decentralized storage network. Identified by a content hash, so any change to the bytes produces a different ID. | | **data availability layer** | Infrastructure guaranteeing a piece of data can be retrieved later by anyone, without depending on a single server. Today's deployment uses Walrus. | | **Merkle root** | A single cryptographic hash that summarizes a set of records. Any change to any record produces a different root, so the root acts as a tamper-detection fingerprint for the whole set. | | {% endtab %} | | {% tab title="Trust and verification" %} | Term | Definition | | --------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **attestation** | A cryptographic statement that something happened the way it is claimed. | | **oracle** | A service that brings off-chain data (price feeds, event outcomes, scheduled triggers) onto a blockchain so that smart contracts can react to it. | | **slashing** | Automatically forfeiting a participant's staked tokens when they violate protocol rules. Slashing only works when the rules being checked are deterministic and market-independent. | | {% endtab %} | | {% tab title="Wallets and operations" %} | Term | Definition | | ---------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | **custodial wallet** | A wallet whose private keys are managed by a service provider rather than the user. | | **BYOI (Bring Your Own Identity)** | Letting partners use any cryptographic identifier they already have (EVM address, Solana keypair, W3C DID) rather than issuing a new one specific to the protocol. | | **testnet** | A non-production deployment of a blockchain for testing. Tokens have no monetary value. | | **mainnet** | The production deployment of a blockchain. Tokens have monetary value and transactions are final. | | **gas** | The fee paid to a blockchain network to process a transaction. | | {% endtab %} | | | {% endtabs %} | | ## Stair AI terminology These terms are concepts native to Glass Box Protocol. Each is defined in detail on its canonical page; the entries below are quick references. {% tabs %} {% tab title="Protocol and scoring" %} | Term | Summary | Canonical page | | ---------------------- | ---------------------------------------------------------------------------------------- | ----------------------------------------------------------------- | | **Glass Box Protocol** | The umbrella name for Stair AI's verifiable-reasoning system. | [Introduction](/stair-ai-docs/readme/introduction.md) | | **Stair AI Score** | Composite 0-100 score of an agent's reasoning quality. Process-only, market-independent. | [Stair AI Score](/stair-ai-docs/concepts/stair-ai-score.md) | | **Reasoning Trace** | An agent's append-only history of records under one `agent_id`. | [The Reasoning Trace](/stair-ai-docs/concepts/reasoning-trace.md) | | {% endtab %} | | | {% tab title="Records and integration" %} | Term | Summary | Canonical page | | --------------------------- | ---------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------- | | **TraceRecord** | Atomic record of one behavior step. Always one word, capitalized. | [The Reasoning Trace](/stair-ai-docs/concepts/reasoning-trace.md) | | **Session** | Group of records sharing a `session_id`, representing one decision cycle. No lifecycle. | [The Reasoning Trace](/stair-ai-docs/concepts/reasoning-trace.md) | | **Three integration modes** | The three ways an agent integrates with Glass Box Protocol. | [Three integration modes](/stair-ai-docs/concepts/integration-modes.md) | | **Framework-native mode** | Highest-trust integration. Agent runs inside a controlled runtime. Label: `framework_controlled`. | [Three integration modes](/stair-ai-docs/concepts/integration-modes.md) | | **SDK-instrumented mode** | Medium-trust integration. Agent runs on own infrastructure, integrates the SDK. Label: `sdk_instrumented`. | [Three integration modes](/stair-ai-docs/concepts/integration-modes.md) | | **Self-attested mode** | Lowest-friction integration. Agent submits traces directly via API. Label: `self_reported`. | [Three integration modes](/stair-ai-docs/concepts/integration-modes.md) | | {% endtab %} | | | | {% endtabs %} | | | ## When to define on a page Three rules for inline Web3-term definitions on any page in this site. 1. Define every Web3 term on its first appearance on the page, even if it has been defined on another page. A reader skimming may land here first. 2. Italicize the term in the prose, then place a hint block immediately after with the formal definition. The hint block should be one or two sentences, not a paragraph. 3. After the first definition on the page, use the term naturally without re-defining it within the same page. Stair-AI-specific terms (Reasoning Trace, TraceRecord, Session, Glass Box Protocol, Stair AI Score, the integration modes) are introduced once on their canonical page and linked from elsewhere. They do not need re-definition on every page.