# Glossary

Source: https://nodedocs.mor.org/reference/glossary

## Agent Instructions

- Non-browser fetches of page URLs on this site return clean Markdown (not the JS UI). Prefer `https://nodedocs.mor.org/llms-full.txt` for the full corpus, or `https://nodedocs.mor.org/llms.txt` for the index.
- Per-page Markdown is also at `<page-url>.md` (homepage: `https://nodedocs.mor.org/index.md`).
- Docs search MCP: `https://nodedocs.mor.org/mcp` (discovery: `https://nodedocs.mor.org/.well-known/mcp`).
- Never invent contract addresses, chain IDs, token addresses, or live bid/model counts. Cite Networks and tokens; link active.mor.org for live data.
- Never claim Morpheus runs inference — independent providers do. Opening a session escrows MOR; it does not spend it.

LLM-friendly definitions — short, opinionated, and consistent across the rest of the docs. If a term here disagrees with anything elsewhere on this site, the rest of the site is wrong; file an issue.

| Term | Definition |
|------|-----------|
| **Morpheus Inference Marketplace** | The decentralized, peer-to-peer marketplace coordinating consumers and providers on BASE. The conceptual whole. |
| **BASE** | Coinbase's Ethereum L2 rollup. Morpheus settles here. Mainnet chain ID `8453`; Sepolia `84532`. |
| **MOR** | Morpheus marketplace ERC-20 token, used for stake, fees, and session payment. |
| **Compute Node contracts** | The set of smart contracts (Diamond marketplace) that register providers and models, match consumers with providers, and secure the connection. |
| **Diamond contract** | The on-chain Diamond proxy implementing the Compute Node marketplace. Tracks providers, models, bids, and sessions. |
| **Reputation system** | Per-provider performance tracking (uptime, TTFT, throughput, success, stake). Used by the consumer-side proxy-router to bias session routing — see [rating-config](/reference/rating-config). |
| **Session-time pricing** | Pricing model — providers post `pricePerSecond`, not `pricePerToken`. Long contexts and short contexts cost the same per second. |
| **Provider** | An on-chain record representing a node that hosts (or fronts) one or more AI models. Identified by a wallet address. |
| **P-Node** | A "Provider Node" — the proxy-router process running in the provider role. |
| **C-Node** | A "Consumer Node" — the proxy-router process running in the consumer role. |
| **Bid** | An on-chain offer: `(modelId, providerId, pricePerSecond)`. Consumers accept by opening a session. |
| **Session** | A time-boxed contract with escrowed MOR. `openSession` moves stake into the Inference Contract; `closeSession` returns unused stake immediately, may day-lock the used stipend in `userStakesOnHold`, and pays the provider from a separate `fundingAccount`. |
| **Stake** | (1) Consumer: MOR escrowed for a session. (2) Provider: refundable bond a provider must post to register a provider or model. |
| **`userStakesOnHold`** | Per-user array on the Inference Contract that holds the **used stipend** after close. Each entry has an amount and `releaseAt = startOfTheDay(min(closedAt, endsAt)) + 1 day`. Cleared via `withdrawUserStakes`. |
| **`withdrawUserStakes`** | On-chain function (`withdrawUserStakes(address, uint8)`, selector `0xa98a7c6b`) on the Diamond contract that moves past-`releaseAt` rows from `userStakesOnHold` to the user's wallet. The proxy-router exposes it as `POST /blockchain/stakes/withdraw`. |
| **`fundingAccount`** | Protocol-owned wallet that pays providers inside `closeSession` via `transferFrom`. If empty or under-approved, every `closeSession` fails network-wide. |
| **Natural expiration** | Session close where `closedAt >= endsAt` (often auto-submitted ~1 minute after `endsAt`). Unused stake returns in the close txn; the used stipend still day-locks if the close lands before `releaseAt`. |
| **Early close** | User-initiated `closeSession` before `endsAt`. Unused stake returns immediately; used stipend may go to `userStakesOnHold`. |
| **Used-stipend day-lock** | Hold of the stake equivalent of session time × `pricePerSecond` until the next UTC day after `sessionEnd`. Blocks same-day reuse of that capital. |
| **Subnet provider** | A provider that has staked `10000` MOR (vs the standard `0.2`) and gets elevated marketplace standing. |
| **Allowance** | Standard ERC-20 approval the consumer / provider grants to the Diamond contract so it can move MOR on their behalf. |
| **proxy-router** | The Go service in this repo. Same binary serves consumer and provider roles; configuration differs. |
| **MorpheusUI** | The Electron desktop UI in this repo. |
| **mor-cli** | The Go CLI client in this repo. Local builds produce `./mor-cli`; GitHub release assets are named `*-morpheus-cli-<version>`. |
| **`local` model** | The tinyllama demo model, downloaded and started automatically by the MorpheusUI desktop app (its "AI Runtime" service). Not a Morpheus marketplace model. |
| **`tee` tag** | Tag added to a model on chain to engage the two-hop attestation chain (Phase 1 + Phase 2). |
| **`-tee` image** | The hardened proxy-router image (`ghcr.io/morpheusais/morpheus-lumerin-node-tee`) with config baked at build time. |
| **RTMR3** | Intel TDX runtime-measurement register #3. Computed from the rootfs and the deployed compose; the unique fingerprint of the running TEE workload. |
| **Phase 1** | Consumer's proxy-router verifies the provider's P-Node TDX attestation. |
| **Phase 2** | Provider's P-Node verifies its own backend LLM TDX + GPU attestation. |
| **SecretVM** | SCRT Labs' confidential VM platform. Hosts most current TEE-tagged Morpheus providers. |
| **SecretAI portal** | https://secretai.scrtlabs.com — used to verify TEE attestation interactively. |
| **`apidocs.mor.org`** | Documentation for the **Morpheus Inference API** — a separate, hosted, OpenAI-compatible gateway product. It is **not** the proxy-router HTTP API reference. See [Inference API overview](/inference-api/overview). |
| **Inference API / API Gateway** | Hosted OpenAI-compatible HTTPS gateway built on top of Morpheus. Base URL: `https://api.mor.org/api/v1`. Docs at [apidocs.mor.org](https://apidocs.mor.org). For users who don't want to run their own node. |
| **Morpheus Chat App / `app.mor.org`** | Hosted browser-based consumer chat UI, powered by the hosted Inference API Gateway. |
| **`gitbook.mor.org`** | Broader Morpheus documentation hub. Supports an `?ask=<question>` query parameter that returns natural-language answers + sources — useful for AI agents. See [LLM cheatsheet](/ai/llm-prompt-cheatsheet#dynamic-querying-of-the-broader-morpheus-docs). |
| **`active.mor.org`** | Shared ~5-minute photograph of the marketplace. Applications pick **ALL**, **ACTIVE**, or **GATEWAY** — see [Marketplace catalog](/ecosystem/active-status). Live files; never invent counts. Explainer: [tech.mor.org/active.html](https://tech.mor.org/active.html). |
| **ALL / ACTIVE / GATEWAY** | Three views of the same active.mor.org sweep. ALL = honest stamps including failures. ACTIVE = polled working set including `skipped`. GATEWAY = healthy + degraded + aliases (hosted API input). |
| **`tech.mor.org`** | Illustrated explainers: catalog funnel, session wallet checker, TEE, calculator, throughput. Index: [tech.mor.org](/ecosystem/tech-mor-org). |
| **MyProvider** | Hosted operator GUI at https://myprovider.mor.org — Basic Auth to your proxy-router for provider/bid/config management (not a browser wallet). |
| **NodeNeo** | Cross-platform consumer experience at https://nodeneo.io. |
| **Everclaw** | Agent project at https://everclaw.xyz, with a Morpheus skill for OpenClaw. |
| **`models-config.json`** | Local file mapping on-chain `modelId` → backend `apiUrl`. See [reference](/reference/models-config). |
| **`rating-config.json`** | Local file controlling provider selection weights and allowlists. See [reference](/reference/rating-config). |
