n8n AI Agent to chat with Airtable and analyze data

Overview

Type: Airtable Workflow
Node Types: lmChatOpenAi, agent, stickyNote, memoryBufferWindow, chatTrigger, executeWorkflowTrigger, set, object, switch, string, aggregate, merge, if, toolWorkflow, toolCode, airtable, httpRequest, body, main, ai_tool, ai_languageModel, ai_memory
Estimated Time: 22 min
Complexity: ⭐⭐⭐ Complex

💬 n8n AI Agent to chat with Airtable and analyze data

Engaging with data stored in Airtable often requires manual navigation and time-consuming searches. This n8n AI Agent to chat with Airtable and analyze data workflow allows users to interact conversationally with their datasets, retrieving essential information quickly while minimizing the need for complex queries.

🔧 Workflow Steps:

  1. When chat message received: Initiates the workflow upon receiving a chat message, acting as the primary trigger for the AI agent.
  2. AI Agent: The core of the workflow, an OpenAI Functions Agent configured as an “Airtable assistant.” It processes user requests, plans and executes tool runs, and can ask clarifying questions.
  3. OpenAI Chat Model: Provides the large language model capabilities that power the AI Agent’s conversational abilities.
  4. Window Buffer Memory: Manages and retains conversation context for the AI Agent, enabling natural, multi-turn dialogues.
  5. Get list of bases: A custom tool workflow that allows the AI Agent to fetch a list of all available Airtable bases, including their IDs and names.
  6. Get base schema: A custom tool workflow that retrieves the schema (tables and their fields with types) for a specified Airtable base.
  7. Search records: A custom tool workflow enabling the AI Agent to search for records within a particular Airtable base and table, with options for filtering, limiting results, specifying fields, and sorting.
  8. Process data with code: A custom tool workflow designed for performing data processing operations that require code, such as mathematical aggregations (e.g., count, sum, average) or generating graphs/images from raw data.
  9. Create map image: A custom code tool that generates a Mapbox image link based on provided geographical coordinates (markers).
  10. Execute Workflow Trigger: Receives commands from the AI Agent (e.g., get_bases, get_base_tables_schema, search, code) to execute the corresponding sub-workflows dynamically.
  11. Switch: Routes the workflow execution to the appropriate branch based on the command received from the “Execute Workflow Trigger” node.
  12. Get Bases: An Airtable node that directly queries the Airtable API to retrieve a list of all available bases.
  13. Get Base/Tables schema: An Airtable node that fetches the detailed schema for a specified Airtable base, outlining its tables and fields.
  14. If filter description exists: Checks if the user’s search request includes a natural language filter description, determining whether to proceed with generating a filter formula.
  15. Set schema and prompt: Prepares the system prompt and JSON schema for an OpenAI API call, guiding it on how to generate valid Airtable filter formulas from natural language.
  16. OpenAI – Generate search filter: Sends the user’s filter description to OpenAI’s chat completion API to generate an Airtable filter formula in the correct format.
  17. Merge: Combines data streams, particularly the generated filter formula with other Airtable search parameters.
  18. Airtable – Search records: An HTTP Request node that constructs and sends a POST request to the Airtable API’s listRecords endpoint, including dynamic parameters for base, table, limits, fields, sorting, and the OpenAI-generated filter formula, handling pagination automatically.
  19. Aggregate, Aggregate1, Aggregate2: These nodes are used to consolidate data from multiple incoming items or to prepare data for the final response.
  20. Response: Prepares and sends the final response back to the user after processing an Airtable query or data aggregation.
  21. OpenAI – Create thread: Initiates a new conversation thread with the OpenAI Assistants API, primarily for code execution and image generation capabilities.
  22. OpenAI – Send message: Sends the user’s request and any relevant data to the OpenAI Assistant thread.
  23. OpenAI – Run assistant: Triggers the OpenAI Assistant to process the message, utilizing its configured tools like the code interpreter.
  24. OpenAI – Get messages: Retrieves the messages from the OpenAI Assistant thread, including any generated outputs or files.
  25. If1: Checks if the retrieved OpenAI message contains file attachments, such as images generated by the code interpreter.
  26. OpenAI – Download File: Downloads any file (e.g., an image) that was generated by the OpenAI Assistant’s code interpreter.
  27. Upload file to get link: Uploads the downloaded file to a temporary file hosting service (tmpfiles.org) to obtain a public, shareable URL for the generated image.
  28. Response1: Sends the public URL of the generated image back to the user.

📌 Use Cases:

  • Retrieve order records, product details, and other relevant data from Airtable through conversational queries.
  • Execute mathematical functions to analyze Airtable data such as calculating averages, sums, and other totals.
  • Optionally generate maps for geographic data visualization directly from Airtable location data.
  • Dynamically query Airtable datasets using natural language prompts.
  • Engage in natural, multi-turn dialogues with the Airtable data, as the agent retains conversation context.
  • Perform tailored searches with specific parameters or filters to refine Airtable results without manual query building.

🧰 Required Credentials:

  • OpenAI API
  • Airtable API
  • Mapbox Public Key

⚙️ Notes & Enhancements:

  • Replace your Mapbox public key in the ‘Create map image’ node’s code.
  • Ensure all OpenAI connections are correctly configured with your credentials.
  • Update all Airtable connections with your account credentials.
  • For better organization and performance, consider separating “Workflow 2” into an additional, dedicated workflow.

Workflow Editor Screenshot

Workflow Screenshot

Workflow JSON Code

{
"nodes": [
{
"id": "799d2e0c-29b9-494c-b11a-d79c7ed4a06d",
"name": "OpenAI Chat Model",
"type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
"position": [
920,
480
],
"parameters": {
"options": {}
},
"credentials": {
"openAiApi": {
"id": "zJhr5piyEwVnWtaI",
"name": "OpenAi club"
}
},
"typeVersion": 1
},
{
"id": "6254ef4e-9699-404e-96a4-569326cce48d",
"name": "AI Agent",
"type": "@n8n/n8n-nodes-langchain.agent",
"position": [
1160,
200
],
"parameters": {
"text": "={{ $('When chat message received').item.json.chatInput }}",
"agent": "openAiFunctionsAgent",
"options": {
"maxIterations": 10,
"systemMessage": "You are Airtable assistant. \nYou need to process user's requests and run relevant tools for that. \n\nPlan and execute in right order runs of tools to get data for user's request.\n\nFeel free to ask questions before do actions - especially if you noticed some inconcistency in user requests that might be error/misspelling. \n\nIMPORTANT Always check right table and base ids before doing queries.\n\nIMPORTANT Use Code function to do aggregation functions that requires math like - count, sum, average and etc. Aggegation function could be recognized by words like \"how many\",\"count\",\"what number\" and etc.\nUse Code function to generate graph and images.\n\nIMPORTANT If search with filter failed - try to fetch records without filter\n\nIMPORTANT Ask yourself before answering - am I did everything is possible? Is the answer is right? Is the answer related to user request?\n\nIMPORTANT Always return in response name of Base and Table where records from. "
},
"promptType": "define"
},
"typeVersion": 1.6
},
{
"id": "227a5427-c270-47dc-bc08-4bb321314926",
"name": "Sticky Note",
"type": "n8n-nodes-base.stickyNote",
"position": [
1740,
620
],
"parameters": {
"height": 80,
"content": "### Replace Mapbox public key - <your_public_key> in code"
},
"typeVersion": 1
},
{
"id": "667751f4-9815-45b7-8dd2-9a0821a7a5a7",
"name": "Sticky Note1",
"type": "n8n-nodes-base.stickyNote",
"position": [
840,
640
],
"parameters": {
"height": 80,
"content": "### Replace OpenAI connection"
},
"typeVersion": 1
},
{
"id": "a9cdec25-4167-44a9-9d3c-fb04aac7bb32",
"name": "Window Buffer Memory",
"type": "@n8n/n8n-nodes-langchain.memoryBufferWindow",
"position": [
1080,
480
],
"parameters": {
"sessionKey": "={{ $('When chat message received').item.json.sessionId }}",
"sessionIdType": "customKey"
},
"typeVersion": 1.3
},
{
"id": "dfab4eb2-ba30-4756-8a52-5d73de9fba53",
"name": "When chat message received",
"type": "@n8n/n8n-nodes-langchain.chatTrigger",
"position": [
940,
200
],
"webhookId": "abf9ab75-eaca-4b91-b3ba-c0f83d3daba4",
"parameters": {
"options": {}
},
"typeVersion": 1.1
},
{
"id": "259e3d13-ca92-4756-af69-34065dbe08f3",
"name": "Execute Workflow Trigger",
"type": "n8n-nodes-base.executeWorkflowTrigger",
"position": [
760,
1340
],
"parameters": {},
"typeVersion": 1
},
{
"id": "5b80c2c8-7649-40f2-b9be-d090d8bd5ae9",
"name": "Response",
"type": "n8n-nodes-base.set",
"position": [
2740,
1360
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "cfdbe2f5-921e-496d-87bd-9c57fdc22a7a",
"name": "response",
"type": "object",
"value": "={{$json}}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "761f5593-f85c-44cd-abbd-aeac78bc31f8",
"name": "Switch",
"type": "n8n-nodes-base.switch",
"position": [
980,
1320
],
"parameters": {
"rules": {
"values": [
{
"outputKey": "get_bases",
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"operator": {
"type": "string",
"operation": "equals"
},
"leftValue": "={{ $('Execute Workflow Trigger').item.json.command }}",
"rightValue": "get_bases"
}
]
},
"renameOutput": true
},
{
"outputKey": "get_base_tables_schema",
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "26a3ffe8-c8a6-4564-8d18-5494a8059372",
"operator": {
"name": "filter.operator.equals",
"type": "string",
"operation": "equals"
},
"leftValue": "={{ $('Execute Workflow Trigger').item.json.command }}",
"rightValue": "get_base_tables_schema"
}
]
},
"renameOutput": true
},
{
"outputKey": "search",
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "0f51cc26-2e42-42e1-a5c2-cb1d2e384962",
"operator": {
"name": "filter.operator.equals",
"type": "string",
"operation": "equals"
},
"leftValue": "={{ $('Execute Workflow Trigger').item.json.command }}",
"rightValue": "search"
}
]
},
"renameOutput": true
},
{
"outputKey": "code",
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "51031140-5ceb-48aa-9f33-d314131a9653",
"operator": {
"name": "filter.operator.equals",
"type": "string",
"operation": "equals"
},
"leftValue": "={{ $('Execute Workflow Trigger').item.json.command }}",
"rightValue": "code"
}
]
},
"renameOutput": true
}
]
},
"options": {}
},
"typeVersion": 3.2
},
{
"id": "d6252c5b-a820-4ded-b59b-ab2fb2e277c3",
"name": "Aggregate",
"type": "n8n-nodes-base.aggregate",
"position": [
1780,
980
],
"parameters": {
"options": {},
"aggregate": "aggregateAllItemData"
},
"typeVersion": 1
},
{
"id": "1442ca2e-1793-4029-b398-61d6e6f1c346",
"name": "Aggregate1",
"type": "n8n-nodes-base.aggregate",
"position": [
1780,
1140
],
"parameters": {
"options": {},
"aggregate": "aggregateAllItemData"
},
"typeVersion": 1
},
{
"id": "a81b4dcc-c999-43be-a0ea-e37f3c7c9f9d",
"name": "Merge",
"type": "n8n-nodes-base.merge",
"position": [
1960,
1360
],
"parameters": {},
"typeVersion": 3
},
{
"id": "8029213c-fd8a-4673-a2a0-11b90fd23971",
"name": "Aggregate2",
"type": "n8n-nodes-base.aggregate",
"position": [
2260,
1360
],
"parameters": {
"options": {
"mergeLists": true
},
"fieldsToAggregate": {
"fieldToAggregate": [
{
"fieldToAggregate": "records"
}
]
}
},
"typeVersion": 1
},
{
"id": "f5f99038-9d19-49ed-9f50-3cd0270bf9ce",
"name": "If1",
"type": "n8n-nodes-base.if",
"position": [
2120,
1720
],
"parameters": {
"options": {},
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "fcb24127-53f9-4498-b0fd-463bd4966ac9",
"operator": {
"type": "string",
"operation": "notExists",
"singleValue": true
},
"leftValue": "={{ $json.data[0].attachments[0].file_id }}",
"rightValue": ""
},
{
"id": "016ecba7-f6af-4881-a7d6-780dcb43223c",
"operator": {
"type": "string",
"operation": "notExists",
"singleValue": true
},
"leftValue": "={{ $json.data[0].content.find(x=>x.type==\"image_file\").image_file.file_id }}",
"rightValue": ""
}
]
}
},
"typeVersion": 2.2
},
{
"id": "abc7ddae-9ca9-4cf6-89a4-a63da8c1e036",
"name": "Response1",
"type": "n8n-nodes-base.set",
"position": [
2760,
1720
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "cfdbe2f5-921e-496d-87bd-9c57fdc22a7a",
"name": "response",
"type": "string",
"value": "={{ $json.data.url.replace('org/','org/dl/') }}"
}
]
},
"includeOtherFields": true
},
"typeVersion": 3.4
},
{
"id": "6f40d50f-70e8-4b64-aa42-ae9262fb8381",
"name": "Sticky Note4",
"type": "n8n-nodes-base.stickyNote",
"position": [
2080,
1520
],
"parameters": {
"width": 160,
"height": 80,
"content": "### Replace Airtable connection"
},
"typeVersion": 1
},
{
"id": "de99a161-5ab3-4b54-bdf7-340d74aa5a93",
"name": "Sticky Note5",
"type": "n8n-nodes-base.stickyNote",
"position": [
1740,
1600
],
"parameters": {
"width": 160,
"height": 80,
"content": "### Replace OpenAI connection"
},
"typeVersion": 1
},
{
"id": "c1e030fd-4449-43ca-a4e7-a863f9487614",
"name": "Sticky Note6",
"type": "n8n-nodes-base.stickyNote",
"position": [
1540,
860
],
"parameters": {
"width": 160,
"height": 80,
"content": "### Replace Airtable connection"
},
"typeVersion": 1
},
{
"id": "4375d3a4-0b3b-4de6-9db7-42af4148af2b",
"name": "Sticky Note3",
"type": "n8n-nodes-base.stickyNote",
"position": [
1360,
1900
],
"parameters": {
"width": 1180,
"height": 80,
"content": "### Replace OpenAI connection"
},
"typeVersion": 1
},
{
"id": "138f813c-d0b0-4a2b-8833-69f1decc9253",
"name": "Sticky Note7",
"type": "n8n-nodes-base.stickyNote",
"position": [
700,
0
],
"parameters": {
"color": 6,
"width": 1320,
"height": 780,
"content": "### Workflow 1"
},
"typeVersion": 1
},
{
"id": "ca87c7b7-ab34-4ff9-8d74-cef90e6f1e5e",
"name": "Sticky Note8",
"type": "n8n-nodes-base.stickyNote",
"position": [
700,
840
],
"parameters": {
"color": 6,
"width": 2240,
"height": 1180,
"content": "### Workflow 2"
},
"typeVersion": 1
},
{
"id": "a5cdf41a-f2ca-4203-94ce-45795395ea92",
"name": "Sticky Note9",
"type": "n8n-nodes-base.stickyNote",
"position": [
300,
680
],
"parameters": {
"color": 7,
"width": 330.5152611046425,
"height": 239.5888196628349,
"content": "### ... or watch set up video [20 min]\n[![Youtube Thumbnail](https://cflobdhpqwnoisuctsoc.supabase.co/storage/v1/object/public/my_storage/Video%2012%20-%20Chat%20with%20Airtable%20Blur.png)](https://youtu.be/SotqsAZEhdc)\n"
},
"typeVersion": 1
},
{
"id": "697889c4-15e7-4099-89b8-f4e2e3a3abac",
"name": "Sticky Note10",
"type": "n8n-nodes-base.stickyNote",
"position": [
0,
0
],
"parameters": {
"color": 7,
"width": 636,
"height": 657,
"content": "![5min Logo](https://cflobdhpqwnoisuctsoc.supabase.co/storage/v1/object/public/my_storage/banner.png)\n## AI Agent to chat with Airtable and analyze data\n**Made by [Mark Shcherbakov](https://www.linkedin.com/in/marklowcoding/) from community [5minAI](https://www.skool.com/5minai)**\n\nEngaging with data stored in Airtable often requires manual navigation and time-consuming searches. This workflow allows users to interact conversationally with their datasets, retrieving essential information quickly while minimizing the need for complex queries.\n\nThis workflow enables an AI agent to facilitate chat interactions over Airtable data. The agent can:\n- Retrieve order records, product details, and other relevant data.\n- Execute mathematical functions to analyze data such as calculating averages and totals.\n- Optionally generate maps for geographic data visualization.\n\n1. **Dynamic Data Retrieval**: The agent uses user prompts to dynamically query the dataset.\n2. **Memory Management**: It retains context during conversations, allowing users to engage in a more natural dialogue.\n3. **Search and Filter Capabilities**: Users can perform tailored searches with specific parameters or filters to refine their results."
},
"typeVersion": 1
},
{
"id": "a9f7c4fd-c07a-4c7c-875d-74b27e3f1fbf",
"name": "Sticky Note11",
"type": "n8n-nodes-base.stickyNote",
"position": [
0,
680
],
"parameters": {
"color": 7,
"width": 280,
"height": 346,
"content": "### Set up steps\n\n1. **Separate workflows**:\n\t- Create additional workflow and move there Workflow 2.\n\n2. **Replace credentials**:\n\t- Replace connections and credentials in all nodes.\n\n3. **Start chat**:\n\t- Ask questions and don't forget to mention required base name."
},
"typeVersion": 1
},
{
"id": "0c86638f-7220-415d-a920-13761da925a6",
"name": "Search records",
"type": "@n8n/n8n-nodes-langchain.toolWorkflow",
"position": [
1500,
480
],
"parameters": {
"name": "search",
"fields": {
"values": [
{
"name": "command",
"stringValue": "search"
}
]
},
"schemaType": "manual",
"workflowId": {
"__rl": true,
"mode": "list",
"value": "zVd0G4m33K6KrBvV",
"cachedResultName": "Airtable Agent Tools"
},
"description": "Search records in specific base and table.\n\n- Use Filter (optional) rules for filtering - describe what logic you want to see in filter including field names. \nIMPORTANT - specify all related fields with types for Filter query with right names based on schema. Tool doesn't know schema and type of fields.\n\n- Use Limit (optional) to get more/less records - default = All records. IMPORTANT use default value only when user ask to get all records for analysis.\n\n- Always try to limit list of fields based on user request or in case of number of fields > 30. IMPORTANT Use fields names only.\n \n- Sort by one/multiple fields if needed - order in array is order of level for sorting.\n\nInput example:\nbase_id - appHwXgLVrBujox4J\ntable_id - tblrGzFneREP5Dktl\nlimit - 100\nsort (optional) - [{\"field\":\"Name\",\"direction\":\"asc\"}]\nfilter_desc (optional) - field Name (string) should be equal/contains Mark\nfields (optional) - [\"Name\",\"Email\"]\n\nOutput example:\nRecord 1 - value 1, value 2",
"inputSchema": "{\n \"type\": \"object\",\n \"properties\": {\n \"base_id\": {\n \"type\": \"string\",\n \"description\": \"ID of the base to search in\"\n },\n \"table_id\": {\n \"type\": \"string\",\n \"description\": \"ID of the table to search in\"\n },\n \"limit\": {\n \"type\": \"number\",\n \"description\": \"Number of records to retrieve (default is all records)\"\n },\n \"filter_desc\": {\n \"type\": \"string\",\n \"description\": \"Text description of the filter logic\"\n },\n \"sort\": {\n \"type\": \"array\",\n \"items\": {\n \"type\": \"object\",\n \"properties\": {\n \"field\": { \"type\": \"string\" },\n \"direction\": { \"type\": \"string\", \"enum\": [\"asc\", \"desc\"] }\n },\n \"required\": [\"field\", \"direction\"]\n },\n \"description\": \"Array of sorting rules for the query\"\n },\n \"fields\": {\n \"type\": \"array\",\n \"items\": { \"type\": \"string\" },\n \"description\": \"List of fields to retrieve\"\n }\n },\n \"required\": [\"base_id\", \"table_id\"]\n}",
"specifyInputSchema": true
},
"typeVersion": 1.2
},
{
"id": "7ba1d6ac-f1a2-4b8d-a9a5-ce92eaa4e7fa",
"name": "Process data with code",
"type": "@n8n/n8n-nodes-langchain.toolWorkflow",
"position": [
1640,
480
],
"parameters": {
"name": "code",
"fields": {
"values": [
{
"name": "command",
"stringValue": "code"
}
]
},
"schemaType": "manual",
"workflowId": {
"__rl": true,
"mode": "list",
"value": "zVd0G4m33K6KrBvV",
"cachedResultName": "Airtable Agent Tools"
},
"description": "Process data with code. Use for math functions and image (graphs) generation. \nIMPORTANT Provide raw data only, don't preprocess or use math functions by yourself\n\nInput example:\nrequest - Count average\ndata - 1,2,3\n\nOutput example:\nAverage is 2\nImage file",
"inputSchema": "{\n \"type\": \"object\",\n \"properties\": {\n \"request\": {\n \"type\": \"string\",\n \"description\": \"Description of the operation to perform.\"\n },\n \"data\": {\n \"type\": \"string\",\n \"description\": \"Stringified data - JSON, strings, arrays and etc.\"\n }\n },\n \"required\": [\"request\", \"data\"]\n}",
"specifyInputSchema": true
},
"typeVersion": 1.2
},
{
"id": "3754175c-6f74-4750-b2e7-00e2bd3caf6d",
"name": "Create map image",
"type": "@n8n/n8n-nodes-langchain.toolCode",
"position": [
1800,
480
],
"parameters": {
"name": "create_map",
"jsCode": "// Example: convert the incoming query to uppercase and return it\n\nreturn `https://api.mapbox.com/styles/v1/mapbox/streets-v12/static/${query.markers}/-96.9749,41.8219,3.31,0/800x500?before_layer=admin-0-boundary&access_token=<your_public_key>`;",
"schemaType": "manual",
"description": "Create link with image for map graph.\nUse addresses' longitude and latitude to create input data.\n\nInput Example:\npin-s+555555(-74.006,40.7128),pin-s+555555(-118.2437,34.0522)\n\nOutput Example:\nImage link.",
"inputSchema": "{\n\"type\": \"object\",\n\"properties\": {\n\t\"markers\": {\n\t\t\"type\": \"string\",\n\t\t\"description\": \"List of markers with longitude and latitude data separated by comma. Keep the same color 555555|Example: pin-s+555555(-74.006,40.7128),pin-s+555555(-118.2437,34.0522)\"\n\t\t}\n\t}\n}",
"specifyInputSchema": true
},
"typeVersion": 1.1
},
{
"id": "135078ea-6a3f-4aee-9f60-c6d5832e446e",
"name": "Get list of bases",
"type": "@n8n/n8n-nodes-langchain.toolWorkflow",
"position": [
1220,
480
],
"parameters": {
"name": "get_bases",
"fields": {
"values": [
{
"name": "command",
"stringValue": "get_bases"
}
]
},
"workflowId": {
"__rl": true,
"mode": "list",
"value": "zVd0G4m33K6KrBvV",
"cachedResultName": "Airtable Agent Tools"
},
"description": "Fetches the list of available bases.\n\nOutput:\n- List of bases with their IDs and names."
},
"typeVersion": 1.2
},
{
"id": "cd4781d0-f873-4aea-951c-6809358c1db6",
"name": "Get base schema",
"type": "@n8n/n8n-nodes-langchain.toolWorkflow",
"position": [
1360,
480
],
"parameters": {
"name": "get_base_tables_schema",
"fields": {
"values": [
{
"name": "command",
"stringValue": "get_base_tables_schema"
}
]
},
"schemaType": "manual",
"workflowId": {
"__rl": true,
"mode": "list",
"value": "zVd0G4m33K6KrBvV",
"cachedResultName": "Airtable Agent Tools"
},
"description": "Fetches the schema of tables in a specific base by id.\n\nInput:\nbase_id: appHwXgLVrBujox4J\n\nOutput:\ntable 1: field 1 - type string, fields 2 - type number",
"inputSchema": "{\n \"type\": \"object\",\n \"properties\": {\n \"base_id\": {\n \"type\": \"string\",\n \"description\": \"ID of the base to retrieve the schema for. Format - appHwXgLVrBujox4J\"\n }\n },\n \"required\": [\"base_id\"]\n}",
"specifyInputSchema": true
},
"typeVersion": 1.2
},
{
"id": "45c8b2eb-f43a-48b1-a270-9caeda9da0b0",
"name": "Get Bases",
"type": "n8n-nodes-base.airtable",
"position": [
1580,
980
],
"parameters": {
"options": {},
"resource": "base"
},
"credentials": {
"airtableTokenApi": {
"id": "xZwG0YpqsxpWrzVM",
"name": "Mark Airtable account"
}
},
"typeVersion": 2.1
},
{
"id": "bb8036bc-1c23-461b-bd03-2461e31c6cb6",
"name": "Get Base/Tables schema",
"type": "n8n-nodes-base.airtable",
"position": [
1580,
1140
],
"parameters": {
"base": {
"__rl": true,
"mode": "id",
"value": "={{ $('Execute Workflow Trigger').item.json.query.base_id }}"
},
"resource": "base",
"operation": "getSchema"
},
"credentials": {
"airtableTokenApi": {
"id": "xZwG0YpqsxpWrzVM",
"name": "Mark Airtable account"
}
},
"typeVersion": 2.1
},
{
"id": "dab309d9-3629-44ba-9f0a-ede55f96488f",
"name": "If filter description exists",
"type": "n8n-nodes-base.if",
"position": [
1340,
1360
],
"parameters": {
"options": {},
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "fcb24127-53f9-4498-b0fd-463bd4966ac9",
"operator": {
"type": "string",
"operation": "notExists",
"singleValue": true
},
"leftValue": "={{ $('Execute Workflow Trigger').item.json.query.filter_desc }}",
"rightValue": ""
}
]
}
},
"typeVersion": 2.2
},
{
"id": "4cc416aa-50bd-4b60-ae51-887c4ee97c88",
"name": "Airtable - Search records",
"type": "n8n-nodes-base.httpRequest",
"onError": "continueErrorOutput",
"position": [
2100,
1360
],
"parameters": {
"url": "=https://api.airtable.com/v0/{{ $('Execute Workflow Trigger').item.json.query.base_id }}/{{ $('Execute Workflow Trigger').item.json.query.table_id }}/listRecords",
"method": "POST",
"options": {
"pagination": {
"pagination": {
"parameters": {
"parameters": [
{
"name": "offset",
"type": "body",
"value": "={{ $response.body.offset}}"
}
]
},
"completeExpression": "={{ $response.body.offset==undefined}}",
"paginationCompleteWhen": "other"
}
}
},
"jsonBody": "={{ \n Object.fromEntries(\n Object.entries({\n sort: $('Execute Workflow Trigger').item.json.query.sort,\n limit: $('Execute Workflow Trigger').item.json.query.limit,\nfields: $('Execute Workflow Trigger').item.json.query.fields,\nfilterByFormula: $('Merge').item.json.choices == undefined ? undefined : JSON.parse($json.choices[0].message.content).filter\n }).filter(([key, value]) => value !== undefined)\n )\n}}",
"sendBody": true,
"specifyBody": "json",
"authentication": "predefinedCredentialType",
"nodeCredentialType": "airtableTokenApi"
},
"credentials": {
"httpQueryAuth": {
"id": "1DXeuNaLSixqGPaU",
"name": "Query Auth account Youtube"
},
"airtableTokenApi": {
"id": "xZwG0YpqsxpWrzVM",
"name": "Mark Airtable account"
}
},
"typeVersion": 4.2
},
{
"id": "9dc71d31-8499-4b69-b87c-898217447d50",
"name": "OpenAI - Generate search filter",
"type": "n8n-nodes-base.httpRequest",
"position": [
1760,
1420
],
"parameters": {
"url": "=https://api.openai.com/v1/chat/completions",
"method": "POST",
"options": {},
"jsonBody": "={\n \"model\": \"gpt-4o-mini\",\n \"messages\": [\n {\n \"role\": \"system\",\n \"content\": {{ JSON.stringify($('Set schema and prompt').item.json.prompt) }}\n },\n {\n \"role\": \"user\",\n \"content\": \"{{ $('Execute Workflow Trigger').item.json.query.filter_desc }}\"\n }],\n \"response_format\":{ \"type\": \"json_schema\", \"json_schema\": {{ $('Set schema and prompt').item.json.schema }}\n\n }\n }",
"sendBody": true,
"specifyBody": "json",
"authentication": "predefinedCredentialType",
"nodeCredentialType": "openAiApi"
},
"credentials": {
"openAiApi": {
"id": "zJhr5piyEwVnWtaI",
"name": "OpenAi club"
}
},
"typeVersion": 4.2
},
{
"id": "16e4ea97-ea73-45a0-aa88-0f9a2969a6a3",
"name": "Set schema and prompt",
"type": "n8n-nodes-base.set",
"position": [
1560,
1420
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "dc09a5b4-ff6a-4cee-b87e-35de7336ac05",
"name": "prompt",
"type": "string",
"value": "=Analyse user request for Airtable filtration. User filter rules to build right formula. Think smart about filter (e.g. instead of search where Name equal to value - search where name contains lowercase value)\nIMPORTANT Check examples and best practices before building formula. \n\nIMPORTANT best practices:\n\nSEARCH(LOWER('example'), LOWER({Field})) ensures both the search term and field are compared in lowercase for consistent case-insensitive matching\n\nIMPORTANT Examples:\n\n- AND(SEARCH('urgent', {Notes}), {Priority} > 3) fetch records where “Notes” contain “urgent” and “Priority” is greater than 3\n- AND({Status} = 'Pending', IS_BEFORE({Due Date}, TODAY())) fetch records where “Status” is “Pending” and “Due Date” is before today\n- OR(SEARCH('error', {Logs}), SEARCH('warning', {Logs})) fetch records where “Logs” contain “error” or “warning”\n- AND(LEN({Description}) > 10, {Price} > 50) fetch records where “Description” is longer than 10 characters and “Price” is greater than 50\n- RECORD_ID() = 'rec12345' fetch a specific record by its ID\n- SEARCH('rec67890', ARRAYJOIN({Linked Records}, ',')) fetch records linked to a specific record ID rec67890\n- AND(SEARCH('rec12345', ARRAYJOIN({Linked Records}, ',')), {Status} = 'Active') fetch records where “Linked Records” contain rec12345 and “Status” is “Active”\n\nFormula rules:\nOperators - =,!=,>,<,>=,<= \n- AND(condition1, condition2, ...) logical AND\n- OR(condition1, condition2, ...) logical OR\n- NOT(condition) logical NOT\n- SEARCH('substring', {Field}) finds position of substring, case-insensitive\n- FIND('substring', {Field}) finds position of substring, case-sensitive\n- IS_BEFORE({Date}, 'YYYY-MM-DD') checks if date is before\n- IS_AFTER({Date}, 'YYYY-MM-DD') checks if date is after\n- IS_SAME({Date1}, {Date2}, 'unit') checks if dates are the same by unit\n- RECORD_ID() = 'recXXXXXX' filters by record ID\n- {Field} = '' field is blank\n- {Field} != '' field is not blank\n- ARRAYJOIN({Linked Field}, ',') joins linked records into a string\n- LOWER({Field}) converts to lowercase for case-insensitive comparison\n- UPPER({Field}) converts to uppercase for case-insensitive comparison\n- VALUE({Text}) converts text to number for numeric comparisons\n- LEN({Field}) gets text length\n- ROUND(number, precision) rounds number\n- TODAY() current date\n- NOW() current timestamp\n- IF(condition, true_value, false_value) conditional logic\n- DATETIME_FORMAT({Date}, 'format') formats date as text\n- DATETIME_DIFF(date1, date2, 'unit') difference between dates\n- DATEADD({Date}, number, 'unit') adds time to date\n- LEFT({Text}, number) extracts leftmost characters\n- RIGHT({Text}, number) extracts rightmost characters\n- AND({Field1} = 'Value1', {Field2} > 50) multiple conditions\n- SEARCH('Value', {Field}) substring match\n- ROUND({Field1} / {Field2}, 2) numeric calculation\n- AND(IS_BEFORE({Date}, TODAY()), {Status} = 'Active') filter by date and status\n- ISERROR(expression) checks if an expression has an error\n- ABS(number) absolute value\n- MIN(value1, value2) minimum value\n- MAX(value1, value2) maximum value\n\n"
},
{
"id": "4e0f9af6-517f-42af-9ced-df0e8a7118b0",
"name": "schema",
"type": "string",
"value": "={\n \"name\": \"filter\",\n \"schema\": {\n \"type\": \"object\",\n \"properties\": {\n \"filter\": {\n \"type\": \"string\"\n }\n },\n \"required\": [\n \"filter\"\n ],\n \"additionalProperties\": false\n },\n \"strict\": true\n}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "6e670074-8508-4282-9c40-600cc445b10f",
"name": "Upload file to get link",
"type": "n8n-nodes-base.httpRequest",
"onError": "continueRegularOutput",
"position": [
2580,
1720
],
"parameters": {
"url": "=https://tmpfiles.org/api/v1/upload",
"method": "POST",
"options": {},
"sendBody": true,
"contentType": "multipart-form-data",
"bodyParameters": {
"parameters": [
{
"name": "file",
"parameterType": "formBinaryData",
"inputDataFieldName": "data"
}
]
}
},
"typeVersion": 4.2
},
{
"id": "b7569d19-3a10-41e5-932b-4be04260a58e",
"name": "OpenAI - Download File",
"type": "n8n-nodes-base.httpRequest",
"position": [
2360,
1720
],
"parameters": {
"url": "=https://api.openai.com/v1/files/{{ $json.data[0].attachments[0]?.file_id ?? $json.data[0].content.find(x=>x.type==\"image_file\")?.image_file.file_id }}/content",
"options": {},
"sendHeaders": true,
"authentication": "predefinedCredentialType",
"headerParameters": {
"parameters": [
{
"name": "OpenAI-Beta",
"value": "assistants=v2"
}
]
},
"nodeCredentialType": "openAiApi"
},
"credentials": {
"openAiApi": {
"id": "vBLHyjEnMK9EaWwQ",
"name": "Mark OpenAi "
}
},
"typeVersion": 4.2
},
{
"id": "bf378b21-07fb-4f9e-bfc5-9623ebcb8236",
"name": "OpenAI - Get messages",
"type": "n8n-nodes-base.httpRequest",
"position": [
1960,
1720
],
"parameters": {
"url": "=https://api.openai.com/v1/threads/{{ $('OpenAI - Create thread').item.json.id }}/messages",
"options": {},
"sendHeaders": true,
"authentication": "predefinedCredentialType",
"headerParameters": {
"parameters": [
{
"name": "OpenAI-Beta",
"value": "assistants=v2"
}
]
},
"nodeCredentialType": "openAiApi"
},
"credentials": {
"openAiApi": {
"id": "zJhr5piyEwVnWtaI",
"name": "OpenAi club"
}
},
"typeVersion": 4.2
},
{
"id": "9874eec1-61e2-45fe-8c57-556957a15473",
"name": "OpenAI - Run assistant",
"type": "n8n-nodes-base.httpRequest",
"position": [
1760,
1720
],
"parameters": {
"url": "=https://api.openai.com/v1/threads/{{ $('OpenAI - Create thread').item.json.id }}/runs",
"method": "POST",
"options": {},
"sendBody": true,
"sendHeaders": true,
"authentication": "predefinedCredentialType",
"bodyParameters": {
"parameters": [
{
"name": "assistant_id",
"value": "asst_PGUuvzEGJWOE8p8vwV56INLO"
},
{
"name": "stream",
"value": "={{true}}"
},
{
"name": "tool_choice",
"value": "={{ {\"type\": \"code_interpreter\"} }}"
},
{
"name": "tools",
"value": "={{ [{\"type\": \"code_interpreter\"}] }}"
}
]
},
"headerParameters": {
"parameters": [
{
"name": "OpenAI-Beta",
"value": "assistants=v2"
}
]
},
"nodeCredentialType": "openAiApi"
},
"credentials": {
"openAiApi": {
"id": "fLfRtaXbR0EVD0pl",
"name": "OpenAi account"
}
},
"typeVersion": 4.2
},
{
"id": "e5339ad2-36c7-40c5-846b-2bd242f41ea5",
"name": "OpenAI - Send message",
"type": "n8n-nodes-base.httpRequest",
"position": [
1560,
1720
],
"parameters": {
"url": "=https://api.openai.com/v1/threads/{{ $('OpenAI - Create thread').item.json.id }}/messages ",
"method": "POST",
"options": {},
"sendBody": true,
"sendHeaders": true,
"authentication": "predefinedCredentialType",
"bodyParameters": {
"parameters": [
{
"name": "role",
"value": "user"
},
{
"name": "content",
"value": "=Request:\n{{ $('Execute Workflow Trigger').item.json.query.request }}\n\nData:\n{{ $('Execute Workflow Trigger').item.json.query.data }}"
}
]
},
"headerParameters": {
"parameters": [
{
"name": "OpenAI-Beta",
"value": "assistants=v2"
}
]
},
"nodeCredentialType": "openAiApi"
},
"credentials": {
"openAiApi": {
"id": "fLfRtaXbR0EVD0pl",
"name": "OpenAi account"
}
},
"typeVersion": 4.2
},
{
"id": "5b822c15-af63-43f6-ac30-61a34dcd91ee",
"name": "OpenAI - Create thread",
"type": "n8n-nodes-base.httpRequest",
"position": [
1360,
1720
],
"parameters": {
"url": "https://api.openai.com/v1/threads",
"method": "POST",
"options": {},
"sendHeaders": true,
"authentication": "predefinedCredentialType",
"headerParameters": {
"parameters": [
{
"name": "OpenAI-Beta",
"value": "assistants=v2"
}
]
},
"nodeCredentialType": "openAiApi"
},
"credentials": {
"openAiApi": {
"id": "vBLHyjEnMK9EaWwQ",
"name": "Mark OpenAi "
}
},
"typeVersion": 4.2
}
],
"pinData": {},
"connections": {
"If1": {
"main": [
[
{
"node": "Response",
"type": "main",
"index": 0
}
],
[
{
"node": "OpenAI - Download File",
"type": "main",
"index": 0
}
]
]
},
"Merge": {
"main": [
[
{
"node": "Airtable - Search records",
"type": "main",
"index": 0
}
]
]
},
"Switch": {
"main": [
[
{
"node": "Get Bases",
"type": "main",
"index": 0
}
],
[
{
"node": "Get Base/Tables schema",
"type": "main",
"index": 0
}
],
[
{
"node": "If filter description exists",
"type": "main",
"index": 0
}
],
[
{
"node": "OpenAI - Create thread",
"type": "main",
"index": 0
}
]
]
},
"Aggregate": {
"main": [
[
{
"node": "Response",
"type": "main",
"index": 0
}
]
]
},
"Get Bases": {
"main": [
[
{
"node": "Aggregate",
"type": "main",
"index": 0
}
]
]
},
"Aggregate1": {
"main": [
[
{
"node": "Response",
"type": "main",
"index": 0
}
]
]
},
"Aggregate2": {
"main": [
[
{
"node": "Response",
"type": "main",
"index": 0
}
]
]
},
"Search records": {
"ai_tool": [
[
{
"node": "AI Agent",
"type": "ai_tool",
"index": 0
}
]
]
},
"Get base schema": {
"ai_tool": [
[
{
"node": "AI Agent",
"type": "ai_tool",
"index": 0
}
]
]
},
"Create map image": {
"ai_tool": [
[
{
"node": "AI Agent",
"type": "ai_tool",
"index": 0
}
]
]
},
"Get list of bases": {
"ai_tool": [
[
{
"node": "AI Agent",
"type": "ai_tool",
"index": 0
}
]
]
},
"OpenAI Chat Model": {
"ai_languageModel": [
[
{
"node": "AI Agent",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Window Buffer Memory": {
"ai_memory": [
[
{
"node": "AI Agent",
"type": "ai_memory",
"index": 0
}
]
]
},
"OpenAI - Get messages": {
"main": [
[
{
"node": "If1",
"type": "main",
"index": 0
}
]
]
},
"OpenAI - Send message": {
"main": [
[
{
"node": "OpenAI - Run assistant",
"type": "main",
"index": 0
}
]
]
},
"Set schema and prompt": {
"main": [
[
{
"node": "OpenAI - Generate search filter",
"type": "main",
"index": 0
}
]
]
},
"Get Base/Tables schema": {
"main": [
[
{
"node": "Aggregate1",
"type": "main",
"index": 0
}
]
]
},
"OpenAI - Create thread": {
"main": [
[
{
"node": "OpenAI - Send message",
"type": "main",
"index": 0
}
]
]
},
"OpenAI - Download File": {
"main": [
[
{
"node": "Upload file to get link",
"type": "main",
"index": 0
}
]
]
},
"OpenAI - Run assistant": {
"main": [
[
{
"node": "OpenAI - Get messages",
"type": "main",
"index": 0
}
]
]
},
"Process data with code": {
"ai_tool": [
[
{
"node": "AI Agent",
"type": "ai_tool",
"index": 0
}
]
]
},
"Upload file to get link": {
"main": [
[
{
"node": "Response1",
"type": "main",
"index": 0
}
]
]
},
"Execute Workflow Trigger": {
"main": [
[
{
"node": "Switch",
"type": "main",
"index": 0
}
]
]
},
"Airtable - Search records": {
"main": [
[
{
"node": "Aggregate2",
"type": "main",
"index": 0
}
],
[
{
"node": "Response",
"type": "main",
"index": 0
}
]
]
},
"When chat message received": {
"main": [
[
{
"node": "AI Agent",
"type": "main",
"index": 0
}
]
]
},
"If filter description exists": {
"main": [
[
{
"node": "Merge",
"type": "main",
"index": 0
}
],
[
{
"node": "Set schema and prompt",
"type": "main",
"index": 0
}
]
]
},
"OpenAI - Generate search filter": {
"main": [
[
{
"node": "Merge",
"type": "main",
"index": 1
}
]
]
}
}
}