Skip to main content

Cognivo-Specific Terms

Definitions for terms used across the Cognivo platform. For blockchain-general terms (CEX, DEX, EOA, smart contracts, bridges, mixers), see the Blockchain Terms page.

$COGNI Gateway​

is the name for Cognivo's built-in blockchain analysis tools. Available directly in Cognivo Chat, these tools let you look up transactions, check wallet activity, analyze smart contracts, view ERC-20 transfers, and trace fund flows — all by asking in natural language. Results link directly to Wallet Tracer for interactive visualization.

Bubble Map​

A is Cognivo Wallet Tracer's primary visualization. It renders an interactive graph where each circle (node) represents an Ethereum address and each arrow (edge) represents a transaction or transfer. Nodes are color-coded by type: regular wallets, exchanges, DEX routers, bridges, and mixers. You can click any node or edge to inspect details, zoom and pan to navigate, and share the map with others.

Terminal​

A address is where a trace path ends. Wallet Tracer stops tracing at an address for specific reasons: the address is a known exchange (CEX_DETECTED), a mixer (MIXER_DETECTED), a bridge (BRIDGE_DETECTED), the maximum trace depth was reached (MAX_DEPTH), the address has no relevant transactions (NO_TRANSACTIONS), or the trace timed out (TIMEOUT). Terminals appear with dashed borders on the Bubble Map.

Terminal Reasons​

explain why Wallet Tracer stopped tracing at a specific address. Each terminal in a trace has a reason badge: CEX_DETECTED (known exchange), MIXER_DETECTED (privacy protocol), BRIDGE_DETECTED (cross-chain bridge), MAX_DEPTH (trace hop limit reached), NO_TRANSACTIONS (address has no relevant outgoing or incoming transactions), or TIMEOUT (trace processing timed out). Understanding terminal reasons helps you interpret trace results accurately.

Hop​

A represents one transaction step in a fund flow trace. When Wallet Tracer follows funds from Address A to Address B, that is one hop. A 5-hop trace can follow funds through up to 5 consecutive transfers. The free tier allows up to 5 hops and 20 addresses per trace.

Node​

On the Bubble Map, a is a circle representing a single Ethereum address. Nodes are color-coded by address type: regular wallets (EOA), exchanges (CEX), DEX routers, bridges, and mixers each have distinct colors. Clicking a node reveals details including the address, type, transaction count, inflow, and outflow amounts.

Edge​

On the Bubble Map, an is a line with an arrow indicating direction that connects two nodes. Each edge represents one or more transactions between two addresses. Clicking an edge shows the transaction hash, value transferred, token type, and direction of the transfer.

Outgoing Trace​

An follows funds forward from the starting address. It answers the question "Where did the money go?" by tracking transfers, swaps, and movements away from the starting wallet through subsequent hops until terminals are reached.

Incoming Trace​

An follows funds backward to the starting address. It answers the question "Where did the money come from?" by looking at who sent funds to the address and tracing those sources further back through the network.

Bidirectional Trace​

A combines both outgoing and incoming traces, following funds in both directions from the starting address. This provides the most complete view of an address's fund flows but may result in more addresses and a more complex Bubble Map.

Multi-Hop Tracing​

is Wallet Tracer's ability to automatically follow funds across multiple consecutive transfers. Instead of just showing who an address sent to directly, it continues following those recipients' transfers, and their recipients' transfers, building a complete picture of fund flows across the network.

Journey View​

The is a tab in Wallet Tracer that presents every transaction in a trace as a chronological, step-by-step list. Each entry shows the sender, receiver, amount, token, and transaction hash. You can click any step to expand it and see full transaction details.

Trace Lifespan​

The tab in Wallet Tracer shows the timeline of a trace's processing. It displays when the trace started, when it finished, and the total elapsed time. This is useful for understanding how long complex traces take to process.

Flow Path Analysis​

is available in the Flow Paths tab of Wallet Tracer. It computes and displays the most significant fund flow routes: top paths ranked by value, the longest path in the trace, categorization of terminal addresses by type, and the most significant counterparty addresses.

Memories​

are pieces of information the AI in Cognivo Chat saves and uses across all your future conversations. You can create memories manually or the AI may create them automatically during a conversation. Memories help the AI provide more relevant responses without you having to repeat background information.

MCP​

(Model Context Protocol) is an open standard that enables AI models to interact with external tools and data sources. In Cognivo Chat, MCP servers extend the AI with access to additional capabilities beyond the built-in tools.

Presets​

in Cognivo Chat are saved configurations that define how the AI behaves. A preset can include the selected model, a system prompt, temperature setting, and other parameters. You can create, save, and switch between presets to quickly change the AI's behavior for different tasks.

Router Address​

A is a DEX smart contract that processes token swaps on behalf of many users. Router addresses are not personal wallets — they are shared infrastructure. In Wallet Tracer, router addresses are identified and labeled separately because their transaction volume and patterns differ fundamentally from personal wallets.

SIWE​

(Sign-In with Ethereum) is an authentication method that lets you log into Cognivo using your Ethereum wallet instead of a traditional email and password. You prove you own the wallet by signing a message — no password is created or stored.

Cognivo Grounded On-Chain Evidence​

Cognivo Grounded On-Chain Evidence is information verified directly from on-chain data. It is the authoritative basis for any on-chain fact in a Cognivo answer. See Grounded vs Augmented Evidence.

Cognivo Augmented Intelligence​

Cognivo Augmented Intelligence is added interpretive context (such as distribution or ranking context) that is always clearly labeled as augmented. It supports interpretation but never overrides grounded evidence.

Cognivo Liquidity Intelligence​

Cognivo Liquidity Intelligence reads liquidity pool context for a token — pair, venue, and approximate liquidity where available — and reports pool coverage honestly. See Liquidity Intelligence.

Burned LP​

Burned LP means liquidity-provider (LP) tokens were sent to a burn address and can no longer be moved. Burned LP has no unlock schedule and is not the same as a timed lock. See Burned-LP Custody.

Timed Lock​

A timed lock restricts withdrawing liquidity for a set period, after which it can be withdrawn. A timed lock is different from burned LP. When Cognivo cannot confirm a timed lock, it reports it as not verified.

Cognivo Registry Attribution​

Cognivo Registry Attribution identifies known infrastructure associated with a token or chain. Registry rows identify known infrastructure only — they are not proof of an active lock state. See Registry Attribution.

Pool Coverage​

Pool coverage is Cognivo's transparent count of how it handled a token's liquidity pools: discovered, checked, readable, and not assessed.

Readable Pool​

A readable pool is a liquidity pool Cognivo could read directly in a given view. Pools that cannot be read the same way are reported as not assessed rather than assumed to be empty.

Not Assessed​

Not assessed means Cognivo did not evaluate a signal in the current view. It is not the same as zero and not a safety claim. See Not Verified, Not Assessed & Augmented.

Not Verified​

Not verified means Cognivo could not confirm a signal. It indicates the absence of confirmation, not that the opposite is true.

Summary Reply​

A summary reply is a shorter, formatted recap of an answer Cognivo already produced, requested with D or summary. It adds no new claims and ends with "No credits were used." when it is a no-credit recap.