@unwarkz/n8n-nodes-assistant-set

Full AI-assistant node set for n8n — Mem0 intelligent memory, Gotenberg PDF conversion, Telegram Bot messaging, Outline Wiki knowledge base, and Qdrant vector store. All nodes include both regular workflow variants and AI Agent tool variants.

Each component is also available as a standalone package:

• @unwarkz/n8n-nodes-mem0
• @unwarkz/n8n-nodes-gotenberg-pdf
• @unwarkz/n8n-nodes-telegram-bot
• @unwarkz/n8n-nodes-outline-wiki
• @unwarkz/n8n-nodes-qdrant

Installation

Via Community Nodes panel (recommended)

1. Go to Settings > Community Nodes and click Install.
2. Enter @unwarkz/n8n-nodes-assistant-set.
3. Confirm and restart n8n.

Manual installation

cd ~/.n8n/nodes
npm install @unwarkz/n8n-nodes-assistant-set

Restart n8n to load the nodes.

Credential Setup

| Credential | How to obtain | |------------|--------------| | Mem0 API | Mem0 Dashboard > API Keys | | Mem0 Self-Hosted API | Your instance URL + optional API key | | Gotenberg API | Your Gotenberg instance URL (default: http://localhost:3000) | | Telegram Bot API | @BotFather — create a bot and copy the token | | Outline API | Your Outline instance URL + API token from Settings > API | | Qdrant API | Your Qdrant instance URL + optional API key |

Included Nodes

Mem0 — Intelligent Memory Layer

| Node | Output | Description | |------|--------|-------------| | Mem0 | main | Full CRUD: add, search, update, delete memories; manage entities, organizations, projects | | Mem0 Memory | ai_memory | Memory source for the AI Agent node — retrieves and saves conversation context | | Mem0 AI Tools | ai_tool | Tools for the AI Agent: mem0_search_memory, mem0_add_memory, mem0_get_all_memories, mem0_delete_memory, mem0_get_memory_history |

Mem0 Memory — Context Retrieval Modes

| Mode | Description | |------|-------------| | Basic | Returns raw stored memories (all or last N) | | Summary | Builds a summarized text block from retrieved memories | | Semantic (v1) | Vector search with query, topK, rerank, fields | | Semantic (v2) | Advanced search with structured filters, topK, rerank, fields | | Hybrid | Combines recent memories + semantic search with time-decay scoring, alpha weighting, and MMR diversity |

The memory node implements loadMemoryVariables to inject context into the agent and saveContext to persist conversation turns back to Mem0 automatically.

Gotenberg — PDF Conversion

Requires a running Gotenberg instance:

docker run --rm -p 3000:3000 gotenberg/gotenberg:8

| Node | Output | Description | |------|--------|-------------| | Gotenberg | main | Convert HTML/URL to PDF, office documents to PDF, screenshots, merge/split/flatten/convert PDFs, read/write metadata | | Gotenberg AI Tools | ai_tool | Same operations as AI Agent tools |

Gotenberg Operations

| Resource | Operation | Description | |----------|-----------|-------------| | Chromium | URL to PDF | Render any public URL as PDF | | Chromium | HTML to PDF | Convert binary HTML to PDF | | Chromium | URL to Screenshot | Take PNG/JPEG/WebP screenshot of a URL | | Chromium | HTML to Screenshot | Screenshot a binary HTML file | | LibreOffice | Convert to PDF | Convert Word, Excel, PowerPoint, ODT, ODS, ODP, CSV, etc. | | PDF Engines | Merge PDFs | Combine multiple PDFs into one | | PDF Engines | Split PDF | Split by page intervals or custom ranges | | PDF Engines | Convert PDF | Convert to PDF/A (PDF/A-1a, PDF/A-2b, PDF/A-3b) | | PDF Engines | Flatten PDF | Flatten annotations and form fields | | PDF Engines | Read Metadata | Read PDF metadata as JSON | | PDF Engines | Write Metadata | Write custom metadata into a PDF |

Telegram Bot — Full Bot API

| Node | Output | Description | |------|--------|-------------| | Telegram Bot | main | 100+ operations covering the complete Telegram Bot API | | Telegram Bot AI Tools | ai_tool | 24 AI Agent tools for messaging, file transfer, and chat management |

Telegram Bot AI Tools

| Tool | Description | |------|-------------| | telegram_send_message | Send text messages with formatting | | telegram_send_photo | Send photos (file_id, URL, or binary property reference) | | telegram_send_document | Send documents/files | | telegram_send_video | Send videos | | telegram_send_audio | Send audio files | | telegram_send_voice | Send voice messages | | telegram_send_location | Send GPS locations | | telegram_send_contact | Send contact cards | | telegram_send_poll | Create polls | | telegram_forward_message | Forward messages between chats | | telegram_edit_message | Edit sent messages | | telegram_delete_message | Delete messages | | telegram_get_file | Download files (stores in binary property for cross-tool use) | | telegram_send_chat_action | Show typing/uploading indicators | | telegram_get_chat | Get chat information | | telegram_send_sticker | Send stickers | | telegram_send_media_group | Send photo/document albums | | telegram_answer_inline_query | Answer inline queries | | telegram_pin_message | Pin messages | | telegram_unpin_message | Unpin messages | | telegram_send_invoice | Send payment invoices | | telegram_get_me | Get bot info | | telegram_set_webhook | Configure webhook | | telegram_delete_webhook | Remove webhook |

Telegram Bot Node — Full API Coverage

• Messages: sendMessage, forwardMessage, copyMessage, sendPhoto, sendAudio, sendDocument, sendVideo, sendAnimation, sendVoice, sendVideoNote, sendMediaGroup, sendLocation, sendVenue, sendContact, sendPoll, sendDice, sendChatAction, editMessageText, editMessageCaption, editMessageMedia, editMessageReplyMarkup, deleteMessage, deleteMessages
• Chat Management: getChat, getChatAdministrators, getChatMemberCount, getChatMember, banChatMember, unbanChatMember, restrictChatMember, promoteChatMember, setChatPermissions, exportChatInviteLink, createChatInviteLink, editChatInviteLink, revokeChatInviteLink, approveChatJoinRequest, declineChatJoinRequest, setChatPhoto, setChatTitle, pinChatMessage, unpinChatMessage, leaveChat
• Bot Info: getMe, logOut, getMyCommands, setMyCommands, deleteMyCommands, getMyName, setMyName, getMyDescription, setMyDescription
• Webhooks: setWebhook, deleteWebhook, getWebhookInfo
• Files: getFile (download with binary data)
• Stickers: sendSticker, getStickerSet, createNewStickerSet, addStickerToSet, deleteStickerFromSet
• Inline: answerInlineQuery, answerWebAppQuery
• Payments: sendInvoice, createInvoiceLink, answerShippingQuery, answerPreCheckoutQuery
• Forum Topics: createForumTopic, editForumTopic, closeForumTopic, reopenForumTopic, deleteForumTopic
• Games: sendGame, setGameScore, getGameHighScores
• Updates: getUpdates

Outline Wiki — Knowledge Base

| Node | Output | Description | |------|--------|-------------| | Outline Wiki | main | Full CRUD for all Outline API resources | | Outline Wiki AI Tools | ai_tool | 31 operations as AI Agent tools |

Outline Wiki — Supported Resources and Operations

| Resource | Operations | |----------|-----------| | Documents | search, searchTitles, create, get, update, delete, list, import (binary), export (binary), archive, restore, move, aiQuestion | | Collections | list, create, get, update, delete, documentTree, export | | Comments | list, create, update, delete | | Attachments | upload (binary → S3 signed URL), delete | | Users | list, get | | Shares | create, list, revoke |

Attachments use a two-step upload: POST /api/attachments.create to obtain a signed S3 URL, then a multipart POST directly to S3.

Qdrant — Vector Store Knowledgebase

| Node | Output | Description | |------|--------|-------------| | Qdrant Store AI Tools | ai_tool | Embed text and binary files into Qdrant using a connected Embedding model sub-node | | Qdrant AI Tools | ai_tool | Full Qdrant operations as AI Agent tools |

Qdrant AI Tools — Operations

| Category | Operations | |----------|-----------| | Search | Semantic search (with optional Reranker sub-node), recommend | | Points | scroll, get, upsert, delete, count | | Collections | list, create, get, update, delete | | Snapshots | create, list, delete |

The Qdrant Store node supports multimodal embedders with automatic base64 fallback for binary files.

Cross-Tool Binary File Interoperability

AI Tools nodes exchange binary files via n8n native binary references — no base64 encoding in the AI context window. Tools that produce files store them via prepareBinaryData and return a binaryPropertyName. Tools that consume files accept a binary_property_name parameter and read data via getBinaryDataBuffer. Compatible with N8N_DEFAULT_BINARY_DATA_MODE=filesystem and database.

Example AI Agent workflows:

Download file from Telegram   →  Convert with Gotenberg         →  Send back via Telegram
Generate PDF from URL         →  Store reference in Mem0        →  Send via Telegram
Receive Telegram attachment   →  Import into Outline Wiki       →  Embed as Outline attachment
Export Outline document       →  Convert to PDF with Gotenberg  →  Send PDF via Telegram
Download image from Telegram  →  Embed with Qdrant Store        →  Search similar images

Node Reference

Mem0 CRUD Node — Parameters

Common Parameters

| Parameter | Description | |-----------|-------------| | Authentication Type | Cloud (api.mem0.ai) or Self-Hosted | | Resource | Memory, Entity, Organization, or Project | | Operation | Depends on resource (Add, Search, Advanced Search, Update, Delete, History, etc.) | | User ID / Agent ID / App ID / Run ID | Scope identifiers for memory operations |

Memory — Advanced Search Filters

Build structured filter rules in the UI:

• Field: e.g., memory, user_id, metadata.category
• Operation: Equals, Not Equals, Contains, Greater Than, Less Than
• Value: string; numbers, booleans, and JSON are parsed automatically
• Rules are combined with AND and sent to the v2 search endpoint

Mem0 Memory — Advanced Options

| Option | Description | |--------|-------------| | User ID / Agent ID / App ID / Run ID | Scope identifiers | | Top K | Number of memories in semantic or hybrid mode (default: 25) | | Rerank / Fields | Available in semantic modes | | Filters (JSON) | Extra filters for Semantic v2 and Hybrid modes | | Last N (recent) | Limit recent memories in Basic, Summary, or Hybrid | | Alpha (semantic weight) | Semantic vs. recency balance in Hybrid (default: 0.65) | | Half-life (hours) | Time-decay half-life in Hybrid (default: 48 h) | | Maximum to Return | Final cap on memories delivered to the AI Agent | | MMR / MMR Lambda | Maximal Marginal Relevance diversity controls | | Save Context | When enabled, conversation turns are persisted to Mem0 automatically |

Self-Hosted Mem0 API Compatibility

| Endpoint | Cloud (api.mem0.ai) | Self-Hosted | |----------|----------------------|-------------| | Semantic search | POST /v1/memories/search/ | POST /search | | Advanced search | POST /v2/memories/search/ | POST /search | | All other endpoints | Same path on both | Same path on both |

Additional self-hosted-only endpoints accessible via the Mem0 node:

• GET /health — service status
• GET /config — current LLM/embedder configuration
• POST /config/switch — switch LLM provider
• POST /reset — reset all memories

Usage Examples

1. Search memories with advanced filters

1. Add a Mem0 node. Set Resource = Memory, Operation = Advanced Search.
2. Enter a query (e.g., "ACME client preferences").
3. Add a filter rule: Field = metadata.segment, Operation = Equals, Value = enterprise.
4. Adjust Top K, Fields, and Rerank as needed.
5. Connect the output to your workflow.

2. Connect Mem0 Memory to an AI Agent (Hybrid mode)

1. Add a Mem0 Memory node and configure credentials.
2. Set Context Retrieval Mode to Hybrid.
3. Set Query to ={{ $json.lastUserMessage }} and Memory Key to chat_history.
4. Adjust Last N, Top K, Alpha, and Half-life as needed.
5. Connect the ai_memory output to the Memory port of the AI Agent node.

3. Convert a URL to PDF with AI Agent

1. Add a Gotenberg AI Tools node. Select the gotenberg_url_to_pdf tool.
2. Connect the ai_tool output to the Tools port of the AI Agent node.
3. The agent can now generate PDFs from URLs on demand.

4. Give the AI Agent Telegram messaging tools

1. Add a Telegram Bot AI Tools node. Configure credentials.
2. Enable the tools you need (e.g., telegram_send_message, telegram_get_file).
3. Connect the ai_tool output to the Tools port of the AI Agent node.

5. Embed documents into Qdrant

1. Add a Qdrant Store AI Tools node. Configure credentials.
2. Connect an Embedding model sub-node to the Embedding input.
3. Connect the ai_tool output to the Tools port of the AI Agent node.
4. The agent can now embed and retrieve documents from Qdrant.

Requirements

• n8n >= 1.0.0
• For Gotenberg nodes: a running Gotenberg instance
• For Mem0 nodes: a Mem0 Cloud account or self-hosted Mem0 API
• For Telegram nodes: a Telegram Bot token from @BotFather
• For Outline nodes: a running Outline instance + API token
• For Qdrant nodes: a running Qdrant instance

License

MIT. Do not commit sensitive information (API keys, personal data) to your workflows or to the npm registry.