# Monte Full LLM Docs # Monte Overview Monte is context personalization for AI agents. The main runtime command is: ```bash monte personalize context "TASK HERE" --json ``` It returns a compact, task-specific agent patch from persona evidence. Agents should apply `instructionBlock`, `personalizedMoves`, `executionProtocol`, `taskAntiPatterns`, and `truthSafety` before doing the work. ## Install ```bash npm i -g monte-engine pnpm add -g monte-engine yarn global add monte-engine bun add -g monte-engine ``` ## Fast path ```bash monte setup monte personalize context "Help me decide what to focus on today without fake progress." --json ``` ## Context modes Use `--mode` when the task type is obvious: - `general`: normal personalized help. - `planning`: plans, sequencing, projects, events, constraints, and follow-through. - `writing`: drafts, tone, messaging, content shape, and audience fit. - `decision`: tradeoffs, recommendations, option scoring, and next moves. - `learning`: explanations, teaching style, practice paths, and comprehension. - `research`: source gathering, evidence synthesis, vendor scans, and comparisons. - `creative`: naming, concepts, campaigns, stories, and divergent options. - `coding` and `design`: explicit preview modes; not auto-selected from broad keywords. Examples: ```bash monte personalize context "Help me approach this task" --mode general --json monte personalize context "Plan next week" --mode planning --json monte personalize context "Draft this update" --mode writing --json monte personalize context "Pick the best option" --mode decision --json monte personalize context "Teach me this topic" --mode learning --json monte personalize context "Research vendor tradeoffs" --mode research --json monte personalize context "Name this product" --mode creative --json ``` ## Sources matter Monte gets useful when it has evidence of how you actually decide, write, prioritize, recover from friction, and work with agents. High-signal sources include chat exports, AI conversations, Obsidian vaults, Notion workspaces, notes, decision logs, project docs, writing samples, task plans, retrospectives, and feedback. Use the guided setup wizard with a folder or export file: ```bash monte setup --files --yes ``` Avoid uploading passwords, API keys, payment details, raw secrets, or files that do not describe how you think or work. ## Profile export is not runtime `monte personalize profile --json` is an inspect/export snapshot, not the main agent runtime. It can be saved locally, but it goes stale, is not task-scoped, lacks live calibration and usage controls, and should not replace fresh `monte personalize context` calls for important work. Use `profile --json` for debugging, inspection, export, or verifying that a persona exists. Use `context` when an agent is about to do real work. ## Links - Website: https://www.monteengine.com/ - Docs: https://www.monteengine.com/docs - LLM map: https://www.monteengine.com/llms.txt - Full LLM docs: https://www.monteengine.com/llms-full.txt --- # Monte CLI Package: `monte-engine` Executable: `monte` ## Install ```bash npm i -g monte-engine pnpm add -g monte-engine yarn global add monte-engine bun add -g monte-engine ``` ## Core commands ```bash monte auth monte auth status monte setup monte setup --files --skill-dir --yes --json monte setup status --json monte skill monte ingest --dry-run monte ingest monte ingest status monte persona build monte persona status monte persona traits monte persona psychology monte personalize bootstrap "" --json monte personalize context "" --json monte personalize profile --json monte personalize calibrate monte doctor ``` ## Hosted API default The hosted CLI already defaults to `https://api.monteengine.com`. Use `monte config set-api` only for local development or alternate deployments. ## Bootstrap `bootstrap` is a lightweight readiness/router command. It tells an agent whether to authenticate, build a persona, calibrate, or call task-aware context next. It is not the main runtime output. ## Context modes Supported auto-classified modes are `general`, `decision`, `writing`, `planning`, `learning`, `research`, and `creative`. `coding` and `design` are explicit preview modes and are not auto-selected from broad task keywords. Examples: ```bash monte personalize context "Help me approach this task" --mode general --json monte personalize context "Plan next week" --mode planning --json monte personalize context "Draft this update" --mode writing --json monte personalize context "Pick the best option" --mode decision --json monte personalize context "Teach me this topic" --mode learning --json monte personalize context "Research vendor tradeoffs" --mode research --json monte personalize context "Name this product" --mode creative --json ``` --- # Monte Sources Sources are an undivided part of Monte. The persona is only as useful as the behavioral evidence it can learn from. ## Useful sources High-signal sources show how you think, decide, write, prioritize, correct mistakes, and work with agents: - ChatGPT, Claude, Codex, Cursor, or other AI conversation exports - Grok/X conversations or X archive data when available - Obsidian vault folders, because they are already Markdown files - Notion Markdown/CSV workspace or page exports - Long-form notes, journals, memos, and planning docs - Decision logs, tradeoff writeups, retrospectives, and postmortems - Project docs, specs, PRDs, roadmaps, and execution plans - Writing samples, drafts, edits, and feedback - Task lists that show follow-through patterns or recurring friction ## What Monte extracts Monte looks for repeated preferences, tradeoff patterns, communication style, friction points, execution habits, uncertainty, corrections, contradictions, and source-specific context. Thin sources create generic personalization. Real sources create sharper context. ## Guided export setup For humans, use the guided wizard: ```bash monte setup ``` With an export file or folder: ```bash monte setup --files --yes ``` For agents, use the async flow: ```bash monte setup --files --skill-dir --yes --json monte setup status --json monte personalize context "TASK HERE" --json ``` ## Source hygiene Prefer plain text, Markdown, JSON, CSV, PDF, Notion Markdown/CSV exports, Obsidian vault folders, or other document folders. Keep filenames meaningful and group sources by app or domain when possible. Do not upload passwords, API keys, payment details, raw secrets, or files that do not describe how you think or work. Hosted raw uploads are temporary processing inputs. Retained evidence is normalized text, signal rows, and evidence packets. ## Official export guides - ChatGPT data export: https://help.openai.com/en/articles/7260999-how-do-i-export-my-data - Claude data export: https://support.claude.com/en/articles/9450526-how-can-i-export-my-claude-data - Google Takeout: https://support.google.com/accounts/answer/3024190 - Notion export: https://www.notion.com/help/export-your-content - Obsidian Markdown/vault guidance: https://help.obsidian.md/import/markdown - X archive: https://help.x.com/managing-your-account/how-to-download-your-twitter-archive Grok is less standardized than ChatGPT or Claude exports. If Grok offers export or data controls in your account, use the exported JSON or Markdown. If you used Grok through X, start with the X archive and include any Grok conversation files it provides. --- # Monte For Agents Use Monte when personalization matters. Skip it for trivial factual or mechanical tasks. ## Agent protocol 1. If Monte is not set up yet, use the guided setup path for humans or the async setup path for agents. 2. For meaningful personalized work, run: ```bash monte personalize context "TASK HERE" --json ``` 3. Read `instructionBlock`, `taskPersonalizationContext.personalizedMoves`, `executionProtocol`, `taskAntiPatterns`, and `truthSafety`. 4. Use explicit modes when clear: `--mode planning`, `--mode writing`, `--mode decision`, `--mode learning`, `--mode research`, or `--mode creative`. 5. Apply the guidance silently. 6. Mention Monte only when auth, credits, persona readiness, staleness, source quality, or calibration blocks the work. 7. Ask at most one blocking clarification question unless the user asked for exploration. ## Reusable skill ```bash monte skill /path/to/agent/skills ``` This creates `/path/to/agent/skills/Monte/SKILL.md`. For agent-safe idempotent setup: ```bash monte skill /path/to/agent/skills --ensure --json ``` ## Copy prompt ```text Use Monte before personalized work. Run: npm i -g monte-engine monte agent setup --key $MONTE_API_KEY --dir . --json monte personalize context "TASK HERE" --json Read the JSON and apply instructionBlock silently. Use Monte for planning, writing, decisions, research, recommendations, creative work, and other tasks where my preferences or constraints matter. Pick an explicit mode when the task type is clear: planning, writing, decision, learning, research, or creative. Skip Monte for trivial facts, simple calculations, and formatting-only edits. Do not use monte personalize profile --json as a replacement for context. Profile JSON is only an inspect/export snapshot. It goes stale, is not task-scoped, lacks live calibration and usage controls, and should not replace fresh context calls for important work. ``` ## Profile export is not runtime `monte personalize profile --json` is for inspection, debugging, and export snapshots. It can be saved, but it goes stale, is not task-scoped, lacks live calibration and usage controls, and should not replace fresh `monte personalize context` calls for important work. ## Simulations Do not run `monte decide` or `monte simulation ...` unless the user explicitly asks for a simulation, scenario test, decision run, or evidence-adjusted rerun. --- # Monte Pricing Hosted Monte uses prepaid wallet credits. Real-persona `monte personalize context` calls cost `$0.20` each. Demo-persona context calls are free up to the demo cap. ## Top-ups - `$10` = 50 real-persona context requests - `$25` = 125 real-persona context requests - `$50` = 250 real-persona context requests Runtime calls return HTTP 402 when real-persona credits are missing.