Documentation / Reference

Glossary

Yuli's own terms, one line each, with a link to the page that explains it.

On this page

Each linked term takes you to the page that explains it in detail. To find one, press / and type its name.

Agent and AI

TermWhat it is
AgentThe AI that plans, proposes changes with their diff and operates the IDE through controlled tools.
Subscription modeYour own Claude Code, Kiro, Codex or Antigravity does the reasoning, with your account.
Platform modeYuli's engine (Yuli Platform) does the reasoning with the API key of the provider you choose, or with a local model.
MCPModel Context Protocol. Yuli includes yuli-mcp, an MCP server that gives your agent the IDE as yuli_* tools.
ACPAgent Client Protocol. Yuli is an ACP client for Kiro, Codex and Antigravity.
Memory, skills and rulesThe agent's lasting context: what it remembers about the project, the procedures it loads and your team's rules.
Agent teamA lead agent splits the work among sub-agents in parallel, integrates the result and has it reviewed.
Unattended modeThe agent asks only while analyzing and carries on by itself until it finishes, without skipping any limit.

Work and control

TermWhat it is
MissionA large piece of work saved with the project: request, requirements, acceptance criteria, decisions and plan.
PhasesThe 8-phase life cycle: analysis, design, planning, implementation, measurement, testing, review and delivery.
AI changesThe window where you review each change from the agent, file by file, before it reaches the disk.
CheckpointThe state each applied task leaves behind: you can go back to it or undo the entire task.
Trusted projectThe decision when you open a project: Trust or Restricted mode, which runs no code from the repository.
SandboxThe perimeter the agent can work in; the engine enforces it on every operation.

Quality and governance

TermWhat it is
N−1The previous version of the code. Yuli measures quality and always compares it with that version: never backwards.
Six SigmaThe defect-measurement framework: DPMO, sigma level, FPY/RTY, Cpk and control chart.
DPMODefects per million opportunities; each line is an opportunity.
Quality contractThe .ai/contract.json file: the team's quality rules, versioned in the repository.
GateThe single command that enforces the contract in Git hooks, CI and the agents.
.ai folderWhere Yuli keeps the project's contract, missions, memory and rules.

Security and privacy

TermWhat it is
MaskingBefore it leaves your computer, a secret or a piece of personal data is replaced by a placeholder with its type and length.
PlaceholderThe text the agent sees instead of the real value. When writing to your project, Yuli restores the value.
Credential vaultAn encrypted area outside the repository where Yuli stores, with your confirmation, loose .env files, keys and tokens.
Super permissionsThe access that lets the agent write to a database; each statement waits for your approval.
Security CenterLocal analysis of dependencies, code, memory leaks and race conditions.

Workspaces

TermWhat it is
ML StudioThe machine learning workspace: data, experiments, training, models and deployment in a guided six-step flow.
LLM StudioThe part of ML Studio that specializes local language models with your examples.
RAGAnswering with your documents: a local index and citations to the source.
Web StudioThe front-end workspace: design, a view of the app, mocked APIs and the full system with Compound.

Account and license

TermWhat it is
Thank youThe page where Yuli thanks the companies and people who support it, everyone equal and in the order they arrived.
Verification codeThe 6 digits sent to your email when you sign up; valid for 15 minutes.
License keyThe free YULI1.… text, signed with Ed25519, that some older BETA builds ask for; it is checked on your computer, offline.
PortableThe Windows executable that opens without installing.
SHA-256 fingerprintThe digest of the file that you compare before opening it, to know it is the original.

See also