AI Travel Agent

# configurations
.idea
.vscode

# crawlee and apify storage folders
apify_storage
crawlee_storage
storage

# installed files
node_modules

# git folder
.git

# dist folder
dist

root = true

[*]
indent_style = space
indent_size = 2
charset = utf-8
trim_trailing_whitespace = true
insert_final_newline = true
end_of_line = lf

USER_APIFY_TOKEN="apify_api_1234"
ACTOR_TEST_PAY_PER_EVENT=true
LANGSMITH_TRACING=false
LANGSMITH_ENDPOINT="https://api.smith.langchain.com"
LANGSMITH_API_KEY="lsv2_pt_1234"
LANGSMITH_PROJECT="langsmith-cool-project-name"
OPENAI_API_KEY="sk-proj-1234"

{
  "root": false,
  "env": {
    "browser": true,
    "es2020": true,
    "node": true
  },
  "extends": [
    "@apify/eslint-config-ts"
  ],
  "parserOptions": {
    "project": "./tsconfig.json",
    "ecmaVersion": 2020
  },
  "ignorePatterns": [
    "node_modules",
    "dist",
    "**/*.d.ts"
  ],
  "rules": {
    "indent": ["error", 2],
    "comma-dangle": ["off", 0],
    "prefer-rest-params": ["off", 0],
    "max-len": ["error", { "ignoreComments": true, "ignoreStrings": true, "ignoreUrls": true, "ignoreTemplateLiterals": true }]
  }
}

# This file tells Git which files shouldn't be added to source control

.idea
.vscode
storage
apify_storage
crawlee_storage
node_modules
dist
tsconfig.tsbuildinfo
.env

# Added by Apify CLI
.venv

image: gitpod/workspace-full:latest
tasks:
  - name: main
    init: >
      nvm use lts/jod &&
      npm install -g npm@11 &&
      npm install -g eslint@^8.50.0 &&
      npm install -g apify-cli
    command: echo "Login to apify to get an API Key, export your APIFY_TOKEN, create an INPUT.json file and run npm start!"
  - name: config
    before: >
      (([[ ! -z $GITCONFIG ]] &&
      echo $GITCONFIG | base64 -d > ~/.gitconfig &&
      chmod 644 ~/.gitconfig) || unset GITCONFIG) &&
      (([[ ! -z $GNUPG_1 ]] &&
      rm -rf ~/.gnupg &&
      cd / &&
      echo $GNUPG_1$GNUPG_2 | base64 -d | tar --no-same-owner -xzf -) || unset GNUPG_1)
    command: echo "Ready! Start a new terminal session in order to use GPG." && exit

Apache License
                           Version 2.0, January 2004
                        http://www.apache.org/licenses/

   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION

   1. Definitions.

      "License" shall mean the terms and conditions for use, reproduction,
      and distribution as defined by Sections 1 through 9 of this document.

      "Licensor" shall mean the copyright owner or entity authorized by
      the copyright owner that is granting the License.

      "Legal Entity" shall mean the union of the acting entity and all
      other entities that control, are controlled by, or are under common
      control with that entity. For the purposes of this definition,
      "control" means (i) the power, direct or indirect, to cause the
      direction or management of such entity, whether by contract or
      otherwise, or (ii) ownership of fifty percent (50%) or more of the
      outstanding shares, or (iii) beneficial ownership of such entity.

      "You" (or "Your") shall mean an individual or Legal Entity
      exercising permissions granted by this License.

      "Source" form shall mean the preferred form for making modifications,
      including but not limited to software source code, documentation
      source, and configuration files.

      "Object" form shall mean any form resulting from mechanical
      transformation or translation of a Source form, including but
      not limited to compiled object code, generated documentation,
      and conversions to other media types.

      "Work" shall mean the work of authorship, whether in Source or
      Object form, made available under the License, as indicated by a
      copyright notice that is included in or attached to the work
      (an example is provided in the Appendix below).

      "Derivative Works" shall mean any work, whether in Source or Object
      form, that is based on (or derived from) the Work and for which the
      editorial revisions, annotations, elaborations, or other modifications
      represent, as a whole, an original work of authorship. For the purposes
      of this License, Derivative Works shall not include works that remain
      separable from, or merely link (or bind by name) to the interfaces of,
      the Work and Derivative Works thereof.

      "Contribution" shall mean any work of authorship, including
      the original version of the Work and any modifications or additions
      to that Work or Derivative Works thereof, that is intentionally
      submitted to Licensor for inclusion in the Work by the copyright owner
      or by an individual or Legal Entity authorized to submit on behalf of
      the copyright owner. For the purposes of this definition, "submitted"
      means any form of electronic, verbal, or written communication sent
      to the Licensor or its representatives, including but not limited to
      communication on electronic mailing lists, source code control systems,
      and issue tracking systems that are managed by, or on behalf of, the
      Licensor for the purpose of discussing and improving the Work, but
      excluding communication that is conspicuously marked or otherwise
      designated in writing by the copyright owner as "Not a Contribution."

      "Contributor" shall mean Licensor and any individual or Legal Entity
      on behalf of whom a Contribution has been received by Licensor and
      subsequently incorporated within the Work.

   2. Grant of Copyright License. Subject to the terms and conditions of
      this License, each Contributor hereby grants to You a perpetual,
      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
      copyright license to reproduce, prepare Derivative Works of,
      publicly display, publicly perform, sublicense, and distribute the
      Work and such Derivative Works in Source or Object form.

   3. Grant of Patent License. Subject to the terms and conditions of
      this License, each Contributor hereby grants to You a perpetual,
      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
      (except as stated in this section) patent license to make, have made,
      use, offer to sell, sell, import, and otherwise transfer the Work,
      where such license applies only to those patent claims licensable
      by such Contributor that are necessarily infringed by their
      Contribution(s) alone or by combination of their Contribution(s)
      with the Work to which such Contribution(s) was submitted. If You
      institute patent litigation against any entity (including a
      cross-claim or counterclaim in a lawsuit) alleging that the Work
      or a Contribution incorporated within the Work constitutes direct
      or contributory patent infringement, then any patent licenses
      granted to You under this License for that Work shall terminate
      as of the date such litigation is filed.

   4. Redistribution. You may reproduce and distribute copies of the
      Work or Derivative Works thereof in any medium, with or without
      modifications, and in Source or Object form, provided that You
      meet the following conditions:

      (a) You must give any other recipients of the Work or
          Derivative Works a copy of this License; and

      (b) You must cause any modified files to carry prominent notices
          stating that You changed the files; and

      (c) You must retain, in the Source form of any Derivative Works
          that You distribute, all copyright, patent, trademark, and
          attribution notices from the Source form of the Work,
          excluding those notices that do not pertain to any part of
          the Derivative Works; and

      (d) If the Work includes a "NOTICE" text file as part of its
          distribution, then any Derivative Works that You distribute must
          include a readable copy of the attribution notices contained
          within such NOTICE file, excluding those notices that do not
          pertain to any part of the Derivative Works, in at least one
          of the following places: within a NOTICE text file distributed
          as part of the Derivative Works; within the Source form or
          documentation, if provided along with the Derivative Works; or,
          within a display generated by the Derivative Works, if and
          wherever such third-party notices normally appear. The contents
          of the NOTICE file are for informational purposes only and
          do not modify the License. You may add Your own attribution
          notices within Derivative Works that You distribute, alongside
          or as an addendum to the NOTICE text from the Work, provided
          that such additional attribution notices cannot be construed
          as modifying the License.

      You may add Your own copyright statement to Your modifications and
      may provide additional or different license terms and conditions
      for use, reproduction, or distribution of Your modifications, or
      for any such Derivative Works as a whole, provided Your use,
      reproduction, and distribution of the Work otherwise complies with
      the conditions stated in this License.

   5. Submission of Contributions. Unless You explicitly state otherwise,
      any Contribution intentionally submitted for inclusion in the Work
      by You to the Licensor shall be under the terms and conditions of
      this License, without any additional terms or conditions.
      Notwithstanding the above, nothing herein shall supersede or modify
      the terms of any separate license agreement you may have executed
      with Licensor regarding such Contributions.

   6. Trademarks. This License does not grant permission to use the trade
      names, trademarks, service marks, or product names of the Licensor,
      except as required for reasonable and customary use in describing the
      origin of the Work and reproducing the content of the NOTICE file.

   7. Disclaimer of Warranty. Unless required by applicable law or
      agreed to in writing, Licensor provides the Work (and each
      Contributor provides its Contributions) on an "AS IS" BASIS,
      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
      implied, including, without limitation, any warranties or conditions
      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
      PARTICULAR PURPOSE. You are solely responsible for determining the
      appropriateness of using or redistributing the Work and assume any
      risks associated with Your exercise of permissions under this License.

   8. Limitation of Liability. In no event and under no legal theory,
      whether in tort (including negligence), contract, or otherwise,
      unless required by applicable law (such as deliberate and grossly
      negligent acts) or agreed to in writing, shall any Contributor be
      liable to You for damages, including any direct, indirect, special,
      incidental, or consequential damages of any character arising as a
      result of this License or out of the use or inability to use the
      Work (including but not limited to damages for loss of goodwill,
      work stoppage, computer failure or malfunction, or any and all
      other commercial damages or losses), even if such Contributor
      has been advised of the possibility of such damages.

   9. Accepting Warranty or Additional Liability. While redistributing
      the Work or Derivative Works thereof, You may choose to offer,
      and charge a fee for, acceptance of support, warranty, indemnity,
      or other liability obligations and/or rights consistent with this
      License. However, in accepting such obligations, You may act only
      on Your own behalf and on Your sole responsibility, not on behalf
      of any other Contributor, and only if You agree to indemnify,
      defend, and hold each Contributor harmless for any liability
      incurred by, or claims asserted against, such Contributor by reason
      of your accepting any such warranty or additional liability.

   END OF TERMS AND CONDITIONS

   APPENDIX: How to apply the Apache License to your work.

      To apply the Apache License to your work, attach the following
      boilerplate notice, with the fields enclosed by brackets "[]"
      replaced with your own identifying information. (Don't include
      the brackets!)  The text should be enclosed in the appropriate
      comment syntax for the file format. We also recommend that a
      file or class name and description of purpose be included on the
      same "printed page" as the copyright notice for easier
      identification within third-party archives.

   Copyright [yyyy] [name of copyright owner]

   Licensed under the Apache License, Version 2.0 (the "License");
   you may not use this file except in compliance with the License.
   You may obtain a copy of the License at

       http://www.apache.org/licenses/LICENSE-2.0

   Unless required by applicable law or agreed to in writing, software
   distributed under the License is distributed on an "AS IS" BASIS,
   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
   See the License for the specific language governing permissions and
   limitations under the License.

1declare global {
2  namespace NodeJS {
3    interface ProcessEnv {
4      USER_APIFY_TOKEN: string;
5      LANGSMITH_TRACING: string;
6      LANGSMITH_ENDPOINT: string;
7      LANGSMITH_API_KEY: string;
8      LANGSMITH_PROJECT: string;
9      OPENAI_API_KEY: string;
10      ZIP_API_KEY: string;
11      NODE_ENV: 'development' | 'production';
12    }
13  }
14}
15
16// If this file has no import/export statements (i.e. is a script)
17// convert it into a module by adding an empty export statement.
18export {}

{
  "name": "ai-travel-agent",
  "version": "0.0.1",
  "type": "module",
  "description": "AI Agent Actor that leverages Apify to search for travel opportunities.",
  "engines": {
    "node": ">=22.0.0"
  },
  "dependencies": {
    "@langchain/core": "^0.3.42",
    "@langchain/langgraph": "^0.2.54",
    "@langchain/openai": "^0.4.4",
    "apify": "^3.3.2",
    "crypto": "^1.0.1",
    "dotenv": "^16.4.7",
    "langchain": "^0.3.19",
    "zod": "^3.24.2"
  },
  "devDependencies": {
    "@apify/eslint-config-ts": "^0.3.0",
    "@apify/tsconfig": "^0.1.0",
    "@eslint/js": "^9.21.0",
    "@types/node": "^22.0.0",
    "@typescript-eslint/eslint-plugin": "^7.18.0",
    "@typescript-eslint/parser": "^7.18.0",
    "eslint": "^8.57.1",
    "globals": "^16.0.0",
    "tsx": "^4.4.0",
    "typescript": "~5.5.0",
    "typescript-eslint": "^8.26.0"
  },
  "scripts": {
    "start": "npm run start:dev",
    "start:prod": "node dist/main.js",
    "start:dev": "tsx src/main.ts",
    "build": "tsc",
    "lint": "eslint src",
    "test": "echo \"Error: oops, the actor has no tests yet, sad!\" && exit 1"
  },
  "author": "It's not you it's me",
  "license": "ISC"
}

{
  "extends": "@apify/tsconfig",
  "compilerOptions": {
    "module": "NodeNext",
    "moduleResolution": "NodeNext",
    "target": "ES2022",
    "outDir": "dist",
    "noUnusedLocals": false,
    "skipLibCheck": true,
    "lib": ["DOM"],
  },
  "include": [
    "./*.d.ts",
    "./src/**/*",
  ]
}

# Specify the base Docker image. You can read more about
# the available images at https://docs.apify.com/sdk/js/docs/guides/docker-images
# You can also use any other image from Docker Hub.
FROM apify/actor-node:22 AS builder

# Check preinstalled packages
RUN npm ls crawlee apify puppeteer playwright

# Copy just package.json and package-lock.json
# to speed up the build using Docker layer cache.
COPY package*.json ./

# Install all dependencies. Don't audit to speed up the installation.
RUN npm install --include=dev --audit=false

# Next, copy the source files using the user set
# in the base image.
COPY . ./

# Install all dependencies and build the project.
# Don't audit to speed up the installation.
RUN npm run build

# Create final image
FROM apify/actor-node:22

# Check preinstalled packages
RUN npm ls crawlee apify puppeteer playwright

# Copy just package.json and package-lock.json
# to speed up the build using Docker layer cache.
COPY package*.json ./

# Install NPM packages, skip optional and development dependencies to
# keep the image small. Avoid logging too much and print the dependency
# tree for debugging
RUN npm --quiet set progress=false \
    && npm install --omit=dev --omit=optional \
    && echo "Installed NPM packages:" \
    && (npm list --omit=dev --all || true) \
    && echo "Node.js version:" \
    && node --version \
    && echo "NPM version:" \
    && npm --version \
    && rm -r ~/.npm

# Copy built JS files from builder image
COPY --from=builder /usr/src/app/dist ./dist

# Next, copy the remaining files and directories with the source code.
# Since we do this after NPM install, quick build will be really fast
# for most source file changes.
COPY . ./

# Create and run as a non-root user.
RUN adduser -h /home/apify -D apify && \
    chown -R apify:apify ./
USER apify

# Run the image.
CMD npm run start:prod --silent

{
  "actorSpecification": 1,
  "name": "ai-travel-agent",
  "version": "0.1",
  "buildTag": "latest",
  "title": "AI Travel Agent",
  "description": "AI agent that uses a multi-agent approach to find you the travel recommendations using a natural language prompt.",
  "meta": {
    "templateId": "ts-empty"
  },
  "environmentVariables": {
    "OPENAI_API_KEY": "@defaultOpenaiApiKey",
    "USER_APIFY_TOKEN": "@defaultApifyToken"
  },
  "dockerfile": "./Dockerfile",
  "input": "./input_schema.json",
  "storages": {
    "dataset": "./dataset_schema.json"
  }
}

{
  "actorSpecification": 1,
  "views": {
    "overview": {
      "title": "Overview",
      "transformation": {
        "fields": ["actorName", "response"]
      },
      "display": {
        "component": "table",
        "properties": {
          "actorName": {
            "label": "Actor name",
            "format": "text"
          },
          "response": {
            "label": "Actor evaluation",
            "format": "text"
          }
        }
      }
    }
  }
}

{
  "title": "Web Automation Agent",
  "type": "object",
  "schemaVersion": 1,
  "properties": {
    "instructions": {
      "title": "Instructions for the AI Travel Agent",
      "type": "string",
      "description": "Ask the agent for help to plan the perfect trip.",
      "editor": "textarea",
      "prefill": "I want to buy a house with a pool in Miami for less than 1 million dollars. Can you help me?"
    },
    "openaiApiKey": {
      "title": "OpenAI API key",
      "type": "string",
      "description": "The API key for accessing OpenAI. You can get it from <a href='https://platform.openai.com/account/api-keys' target='_blank' rel='noopener'>OpenAI platform</a>.",
      "editor": "textfield",
      "isSecret": true
    },
    "model": {
      "title": "GPT model",
      "type": "string",
      "description": "Select a GPT model. See <a href='https://platform.openai.com/docs/models/overview' target='_blank' rel='noopener'>models overview</a>. Keep in mind that each model has different pricing and features.",
      "editor": "select",
      "default": "gpt-4o-mini",
      "prefill": "gpt-4o-mini",
      "enum": ["gpt-4o", "gpt-4o-mini", "gpt-3.5-turbo"]
    },
    "debug": {
      "title": "Debug Mode",
      "type": "boolean",
      "description": "Mark this option as true if you want to see all the logs when running the actor.",
      "editor": "checkbox",
      "default": false
    }
  },
  "required": ["instructions", "model"]
}

{
  "actor-start-gb": {
    "eventTitle": "Actor start per 1 GB",
    "eventDescription": "Flat fee for starting an Actor run for each 1 GB of memory.",
    "eventPriceUsd": 0.01
  },
  "openai-1000-tokens-gpt-4o": {
    "eventTitle": "Price per 1000 OpenAI tokens for gpt-4o",
    "eventDescription": "Flat fee for every 1000 tokens (input/output) used with gpt-4o.",
    "eventPriceUsd": 0.01
  },
  "openai-1000-tokens-gpt-4o-mini": {
    "eventTitle": "Price per 1000 OpenAI tokens for gpt-4o-mini",
    "eventDescription": "Flat fee for every 1000 tokens (input/output) used with gpt-4o-mini.",
    "eventPriceUsd": 0.0006
  },
  "openai-1000-tokens-gpt-3-5-turbo": {
    "eventTitle": "Price per 1000 OpenAI tokens for gpt-3.5-turbo",
    "eventDescription": "Flat fee for every 1000 tokens (input/output) used with gpt-3.5-turbo.",
    "eventPriceUsd": 0.0015
  },
  "duck-duck-go": {
    "eventTitle": "Price per Duck Duck Go search",
    "eventDescription": "Flat fee for every Duck Duck Go search.",
    "eventPriceUsd": 0.01
  },
  "website-scraper':": {
    "eventTitle": "Price per web page scraped",
    "eventDescription": "Flat fee for every web page scraped.",
    "eventPriceUsd": 0.01
  },
  "airbnb-search':": {
    "eventTitle": "Price per Airbnb search result",
    "eventDescription": "Flat fee for every result when searching Airbnb.",
    "eventPriceUsd": 0.001
  },
  "tripadvisor-search':": {
    "eventTitle": "Price per TripAdvisor search result",
    "eventDescription": "Flat fee for every result when searching TripAdvisor.",
    "eventPriceUsd": 0.003
  }
}

1import dotenv from 'dotenv';
2import { Actor, ApifyClient, log } from 'apify';
3import { StateGraph, StateGraphArgs, START, END } from '@langchain/langgraph';
4import LocationAgent from './agents/location.js';
5import DeciderAgent from './agents/decider.js';
6import ResearcherAgent from './agents/researcher.js';
7import SuccessAgent from './agents/success.js';
8import { chargeForActorStart } from './utils/ppe_handler.js';
9
10// if available, .env variables are loaded
11dotenv.config();
12
13/**
14 * Actor input schema
15 */
16type Input = {
17  instructions: string;
18  modelName?: string;
19  openaiApiKey?: string;
20  debug?: boolean;
21}
22
23/**
24 * Actor initialization code and initial charge
25*/
26await Actor.init();
27await chargeForActorStart();
28
29// Handle and validate input
30const {
31  instructions,
32  modelName = 'gpt-4o-mini',
33  openaiApiKey,
34  debug,
35} = await Actor.getInput() as Input;
36if (debug) {
37  log.setLevel(log.LEVELS.DEBUG);
38} else {
39  log.setLevel(log.LEVELS.INFO);
40}
41// if available, the Actor uses the user's openaiApiKey. Otherwise it charges for use.
42const tokenCostActive = (openaiApiKey ?? '').length === 0;
43if (tokenCostActive) {
44  log.info("No openaiApiKey was detected. You'll be charged for token usage.");
45} else {
46  log.info("Env openaiApiKey detected. You won't be charged for token usage.");
47}
48if (!instructions) {
49  throw new Error('Instructions are required. Create an INPUT.json file in the `storage/key_value_stores/default` folder and add the respective keys.');
50}
51
52// Apify is used to call tools and manage datasets
53const userToken = process.env.USER_APIFY_TOKEN;
54if (!userToken) {
55  throw new Error('User token is required. Export your Apify secret as USER_APIFY_TOKEN.');
56}
57const apifyClient = new ApifyClient({
58  token: userToken,
59});
60
61/**
62 * LangGraph StateGraph schema
63 */
64type StateSchema = {
65  instructions: string;
66  bestLocations: string;
67  datasetId: string;
68  totalItems: number;
69  assumptions: string;
70  itemsChecked: number;
71  recommendations: string[];
72  output: string;
73}
74
75const graphState: StateGraphArgs<StateSchema>['channels'] = {
76  instructions: {
77    value: (x?: string, y?: string) => y ?? x ?? '',
78    default: () => instructions,
79  },
80  bestLocations: {
81    value: (x?: string, y?: string) => y ?? x ?? '',
82    default: () => '',
83  },
84  datasetId: {
85    value: (x?: string, y?: string) => y ?? x ?? '',
86    default: () => '',
87  },
88  totalItems: {
89    value: (x?: number, y?: number) => y ?? x ?? 0,
90    default: () => 0,
91  },
92  assumptions: {
93    value: (x?: string, y?: string) => y ?? x ?? '',
94    default: () => '',
95  },
96  itemsChecked: {
97    value: (x?: number, y?: number) => y ?? x ?? 0,
98    default: () => 0,
99  },
100  recommendations: {
101    value: (x?: string[], y?: string[]) => y ?? x ?? [],
102    default: () => [],
103  },
104  output: {
105    value: (x?: string, y?: string) => y ?? x ?? '',
106    default: () => '',
107  },
108};
109
110async function locationNode(state: StateSchema) {
111  const locationAgent = new LocationAgent({
112    apifyClient,
113    modelName,
114    openaiApiKey: openaiApiKey ?? process.env.OPENAI_API_KEY,
115    log,
116  });
117  const { agentExecutor, costHandler } = locationAgent;
118  const response = await agentExecutor.invoke({ input: state.instructions });
119  log.debug(`locationAgent 🤖 : ${response.output}`);
120  await costHandler.logOrChargeForTokens(modelName, tokenCostActive);
121  return { bestLocations: response.output };
122}
123
124async function researcherNode(state: StateSchema) {
125  const researcherAgent = new ResearcherAgent({
126    apifyClient,
127    modelName,
128    openaiApiKey: openaiApiKey ?? process.env.OPENAI_API_KEY,
129    log,
130  });
131  const { agentExecutor, costHandler } = researcherAgent;
132  const input = 'The user asked sent this exact query:\n\n'
133    + `'${state.instructions}'\n\n`
134    + 'You asked someone for help to get the best locations in that city and country. '
135    + 'The answer you received was this:\n\n'
136    + `'${state.bestLocations}'\n\n`
137    + 'Please make some research to gather information to be able to answer the user accordingly.';
138  const response = await agentExecutor.invoke({ input });
139  log.debug(`researcherAgent 🤖 : ${response.output}`);
140  const { datasetId, totalItems, assumptions } = JSON.parse(response.output);
141  await costHandler.logOrChargeForTokens(modelName, tokenCostActive);
142  return { datasetId, totalItems, assumptions };
143}
144
145async function deciderNode(state: StateSchema) {
146  const deciderAgent = new DeciderAgent({
147    apifyClient,
148    modelName,
149    openaiApiKey: openaiApiKey ?? process.env.OPENAI_API_KEY,
150    log,
151  });
152  const { agentExecutor, costHandler } = deciderAgent;
153  const input = 'The user asked sent this exact query:\n\n'
154    + `'${state.instructions}'\n\n`
155    + 'You asked someone for help to get the best locations in that city and country. '
156    + 'The answer you received was this:\n\n'
157    + `'${state.bestLocations}'\n\n`
158    + 'You asked someone else for help to get the best places to stay in in those locations using Airbnb. '
159    + 'The answer you received was this:\n\n'
160    + `Airbnb Apify dataset ID: '${state.datasetId}'\n`
161    + `Total items in Airbnb dataset: '${state.totalItems}'\n`
162    + `Assumptions made when cretating the Airbnb dataset:: '${state.assumptions}'\n`
163    + `Total results already checked in Airbnb dataset: itemsChecked='${state.itemsChecked}'\n\n`
164    + 'Please explore theavailable datasets for the best results.';
165  const response = await agentExecutor.invoke({ input });
166  log.debug(`deciderAgent 🤖 : ${response.output}`);
167  const { itemsChecked, recommendations } = JSON.parse(response.output);
168  await costHandler.logOrChargeForTokens(modelName, tokenCostActive);
169  return {
170    itemsChecked: state.itemsChecked + itemsChecked,
171    recommendations: [...state.recommendations, ...recommendations]
172  };
173}
174
175const resultsCheckedRouter = (state: StateSchema) => {
176  const allResultsChecked = state.itemsChecked >= state.totalItems;
177  log.debug(`allResultsChecked: ${allResultsChecked}`);
178  const thousandResultsChecked = state.itemsChecked > 1000;
179  log.debug(`thousandResultsChecked: ${thousandResultsChecked}`);
180  const fifteenRecommendationsFound = state.recommendations.length > 15;
181  log.debug(`fifteenRecommendationsFound: ${fifteenRecommendationsFound}`);
182  if (
183    allResultsChecked
184    || thousandResultsChecked
185    || fifteenRecommendationsFound
186  ) {
187    // if any of the above criteria are met, it passes the ball to the successAgent
188    return 'success';
189  }
190  // if none of the above criteria are met, it runs the same node again
191  return 'decider';
192};
193
194async function successNode(state: StateSchema) {
195  const successAgent = new SuccessAgent({
196    apifyClient,
197    modelName,
198    openaiApiKey: openaiApiKey ?? process.env.OPENAI_API_KEY,
199    log,
200  });
201  const { agentExecutor, costHandler } = successAgent;
202  const input = 'The user asked sent this exact query:\n\n'
203    + `'${state.instructions}'\n\n`
204    + 'You asked someone for help to get the best locations in that city and state. '
205    + 'The answer you received was this:\n\n'
206    + `'${state.bestLocations}'\n\n`
207    + 'You asked someone else for help to get the best places to stay in those locations using Airbnb and their expert jugdgement. '
208    + 'The answer you received was this:\n\n'
209    + `Assumptions made when searching Airbnb: '${state.assumptions}'\n\n`
210    + `Total places to stay checked: '${state.itemsChecked}'\n`
211    + `Total places to stay recommended: '${state.recommendations.length}'\n`
212    + `Best places to stay in stringified JSON format: '${JSON.stringify(state.recommendations)}'\n`
213    + 'Please select the top 5 places to stay and answer the user explaining the whole process in markdown format.';
214  const response = await agentExecutor.invoke({ input });
215  log.debug(`successNode 🤖 : ${response.output}`);
216  await costHandler.logOrChargeForTokens(modelName, tokenCostActive);
217  return { output: response.output };
218}
219
220const graph = new StateGraph({ channels: graphState })
221  .addNode('location', locationNode)
222  .addNode('researcher', researcherNode)
223  .addNode('decider', deciderNode)
224  .addNode('success', successNode)
225  .addEdge(START, 'location')
226  .addEdge('location', 'researcher')
227  .addEdge('researcher', 'decider')
228  .addConditionalEdges('decider', resultsCheckedRouter)
229  .addEdge('success', END);
230
231const runnable = graph.compile();
232
233const response = await runnable.invoke(
234  { input: instructions },
235  { configurable: { thread_id: 42 } }, // this line shows that the agent can be thread-aware
236);
237
238log.debug(`Agent 🤖 : ${response.output}`);
239
240await Actor.pushData({
241  actorName: 'AI Travel Agent',
242  response: response.output,
243});
244
245await Actor.exit();

1import { Log, ApifyClient } from 'apify';
2import { ChatPromptTemplate } from '@langchain/core/prompts';
3import { createToolCallingAgent, AgentExecutor } from 'langchain/agents';
4import { ChatOpenAI } from '@langchain/openai';
5import { StructuredToolInterface } from '@langchain/core/tools';
6import DatasetExplorer from '../tools/dataset_explorer.js';
7import { CostHandler } from '../utils/cost_handler.js';
8
9/**
10 * Interface for parameters required by PropertyAgent class.
11 */
12export interface PropertyAgentParams {
13  apifyClient: ApifyClient,
14  modelName: string,
15  openaiApiKey: string,
16  log: Log | Console;
17}
18
19/**
20 * An AI Agent that explores an Apify dataset in search for the perfect results.
21 */
22export class PropertyAgent {
23  protected log: Log | Console;
24  protected apifyClient: ApifyClient;
25  public agentExecutor: AgentExecutor;
26  public costHandler: CostHandler;
27
28  constructor(fields?: PropertyAgentParams) {
29    this.log = fields?.log ?? console;
30    this.apifyClient = fields?.apifyClient ?? new ApifyClient();
31    this.costHandler = new CostHandler(fields?.modelName ?? 'gpt-4o-mini', this.log);
32    const llm = new ChatOpenAI({
33      model: fields?.modelName,
34      apiKey: fields?.openaiApiKey,
35      temperature: 0,
36      callbacks: [
37        this.costHandler,
38      ],
39    });
40    const tools = this.buildTools(this.apifyClient, this.log);
41    const prompt = this.buildPrompt();
42    const agent = createToolCallingAgent({
43      llm,
44      tools,
45      prompt,
46    });
47    this.agentExecutor = new AgentExecutor({
48      agent,
49      tools,
50      verbose: false,
51      maxIterations: 5,
52    });
53  }
54
55  protected buildTools(
56    apifyClient: ApifyClient, log: Log | Console
57  ): StructuredToolInterface[] {
58    // Tools are initialized to be passed to the agent
59    const datasetExplorer = new DatasetExplorer({ apifyClient, log });
60    return [
61      datasetExplorer,
62    ];
63  }
64
65  protected buildPrompt(): ChatPromptTemplate {
66    return ChatPromptTemplate.fromMessages([
67      ['system',
68        'You are a experienced travel agent that wants to help the user plan the perfect trip. '
69        + "You'll have a datasetId and the total amount of items available to explore. "
70        + 'Explore only one batch of 25 results (using limit=25). '
71        + 'Skip the first items based on the number of items that were already checked (using offset=itemsChecked). '
72        + "When exploring the dataset, include only these fields exactly as written here: ['id','name','title','rating','pricing','url','images']"
73        + 'Remember: do not call the DatasetExplorer tool more than once, or you will die. \n\n'
74        + 'The user may have specified other requests like pets, bedrooms, beds, baths, etc. '
75        + 'Filter the results based on this information or notify the user if you are unable to do so. '
76        + 'Use your expertise to recommend the best results that you can find that matches the user criteria. '
77        + 'If the total amount of items (totalItems) is over 100, select up to 2 results otherwise pick up to 3. Try to at least pick one. '
78        + "As your response, return only a JSON object (and nothing more! not even a ```json or ``` wrapper) with the key 'itemsChecked' as a number with the amount of results that you received from the dataset_explorer and the key 'recommendations' with your recommendations in JSON format, using the same fields that you specified when calling the dataset_explorer."
79      ],
80      ['placeholder', '{chat_history}'],
81      ['human', '{input}'],
82      ['placeholder', '{agent_scratchpad}'],
83    ]);
84  }
85}
86
87export default PropertyAgent;

1import { Log, ApifyClient } from 'apify';
2import { ChatPromptTemplate } from '@langchain/core/prompts';
3import { createToolCallingAgent, AgentExecutor } from 'langchain/agents';
4import { ChatOpenAI } from '@langchain/openai';
5import { StructuredToolInterface } from '@langchain/core/tools';
6import DuckDuckGo from '../tools/duck_duck_go.js';
7import WebsiteScraper from '../tools/website_scraper.js';
8import { CostHandler } from '../utils/cost_handler.js';
9
10/**
11 * Interface for parameters required by LocationAgent class.
12 */
13export interface LocationAgentParams {
14  apifyClient: ApifyClient,
15  modelName: string,
16  openaiApiKey: string,
17  log: Log | Console;
18}
19
20/**
21 * AI Agent that takes the user input and finds the best Zip Codes in a city to live in.
22 */
23export class LocationAgent {
24  protected log: Log | Console;
25  protected apifyClient: ApifyClient;
26  public agentExecutor: AgentExecutor;
27  public costHandler: CostHandler;
28
29  constructor(fields?: LocationAgentParams) {
30    this.log = fields?.log ?? console;
31    this.apifyClient = fields?.apifyClient ?? new ApifyClient();
32    this.costHandler = new CostHandler(fields?.modelName ?? 'gpt-4o-mini', this.log);
33    const llm = new ChatOpenAI({
34      model: fields?.modelName,
35      apiKey: fields?.openaiApiKey,
36      temperature: 0,
37      callbacks: [
38        this.costHandler,
39      ],
40    });
41    const tools = this.buildTools(this.apifyClient, this.log);
42    const prompt = this.buildPrompt();
43    const agent = createToolCallingAgent({
44      llm,
45      tools,
46      prompt,
47    });
48    this.agentExecutor = new AgentExecutor({
49      agent,
50      tools,
51      verbose: false,
52      maxIterations: 5,
53    });
54  }
55
56  protected buildTools(
57    apifyClient: ApifyClient, log: Log | Console
58  ): StructuredToolInterface[] {
59    // Tools are initialized to be passed to the agent
60    const duckDuckGo = new DuckDuckGo({ apifyClient, log });
61    const websiteScraper = new WebsiteScraper({ apifyClient, log });
62    return [
63      duckDuckGo,
64      websiteScraper,
65    ];
66  }
67
68  protected buildPrompt(): ChatPromptTemplate {
69    return ChatPromptTemplate.fromMessages([
70      ['system',
71        'You are a experienced travel agent that wants to help the user plan their trip. '
72        + "You will stick to the following steps, ignoring the specifics of the user's query and focusing only on finding the best locations. "
73        + 'Step 1. The user will ask you for advice regarding traveling to a specific city in the world or even a neighborhood in a city. '
74        + "You don't care where the user lives, since you are not in charge of flights. "
75        + 'If the user does not specify a budget, select mid range (and not low budget or luxury). '
76        + 'If the user does not provide a country, try to guess to which country the city belongs to. '
77        + 'If from the input you cannot get a city and a country or you think that the specified city should not be visited by tourists, '
78        + 'end the conversation and help the user with the input. '
79        // 2. Fetch best neighborhoods to stay at using DuckDuckGo
80        + 'Step 2. With the city and state in hand, use DuckDuckGo replacing the following query: '
81        + "'where to stay and where not to stay in [city], [country]'. "
82        + 'If there the results, use a website scraper on the URLs you think will show you this information. '
83        + 'Iterate until you find the best 3 locations or neighborhoods in the selected city. '
84        + "For best performance when using the website scraper, you can use method='getBestPlacesToStay' and output=null'. "
85        + 'Instead of answering the original question, just return the top 3 locations or neighborhoods and explain to the user why you selected those locations or neighborhoods. \n'
86      ],
87      ['placeholder', '{chat_history}'],
88      ['human', '{input}'],
89      ['placeholder', '{agent_scratchpad}'],
90    ]);
91  }
92}
93
94export default LocationAgent;

1import { Log, ApifyClient } from 'apify';
2import { ChatPromptTemplate } from '@langchain/core/prompts';
3import { createToolCallingAgent, AgentExecutor } from 'langchain/agents';
4import { ChatOpenAI } from '@langchain/openai';
5import { StructuredToolInterface } from '@langchain/core/tools';
6import AirbnbSearch from '../tools/airbnb_search.js';
7import { CostHandler } from '../utils/cost_handler.js';
8
9/**
10 * Interface for parameters required by ResearcherAgent class.
11 */
12export interface ResearcherAgentParams {
13  apifyClient: ApifyClient,
14  modelName: string,
15  openaiApiKey: string,
16  log: Log | Console;
17}
18
19/**
20 * An AI Agent that searches Researcher for specific locations and stores the results in a dataset.
21 */
22export class ResearcherAgent {
23  protected log: Log | Console;
24  protected apifyClient: ApifyClient;
25  public agentExecutor: AgentExecutor;
26  public costHandler: CostHandler;
27
28  constructor(fields?: ResearcherAgentParams) {
29    this.log = fields?.log ?? console;
30    this.apifyClient = fields?.apifyClient ?? new ApifyClient();
31    this.costHandler = new CostHandler(fields?.modelName ?? 'gpt-4o-mini', this.log);
32    const llm = new ChatOpenAI({
33      model: fields?.modelName,
34      apiKey: fields?.openaiApiKey,
35      temperature: 0,
36      callbacks: [
37        this.costHandler,
38      ],
39    });
40    const tools = this.buildTools(this.apifyClient, this.log);
41    const prompt = this.buildPrompt();
42    const agent = createToolCallingAgent({
43      llm,
44      tools,
45      prompt,
46    });
47    this.agentExecutor = new AgentExecutor({
48      agent,
49      tools,
50      verbose: false,
51      maxIterations: 3,
52    });
53  }
54
55  protected buildTools(
56    apifyClient: ApifyClient, log: Log | Console
57  ): StructuredToolInterface[] {
58    // Tools are initialized to be passed to the agent
59    const airbnbSearch = new AirbnbSearch({ apifyClient, log });
60    return [
61      airbnbSearch,
62    ];
63  }
64
65  protected buildPrompt(): ChatPromptTemplate {
66    const today = new Date().toISOString().slice(0, 10);
67    return ChatPromptTemplate.fromMessages([
68      ['system',
69        'You are a experienced travel agent that knows that Airbnb is a website where you can search for places to stay based on criteria like location, dates and price. '
70        + 'Using the recommended locations, search on Airbnb for places that match the user price, date and travelers (adults, children and pets) requirements. '
71        + 'The user must specify who is traveling (adults, children, infants, pets) (default to 2 adults if the information is not given). '
72        + `When managing dates, you always try to use the best format in the world, which is: yyy-mm-dd. For any calculation assume that today is ${today}.`
73        + 'The user can specify if the travel has a start date (default to a trip of one week if the information is not given and choose today if no start date is available). '
74        + 'The user can specify if the travel has an end date (default to a trip of one week if the information is not given and choose today if no start date is available). '
75        + 'The user can specify if the travel destination has a minimum value (default to 1 if the information is not given). '
76        + 'The user can specify if the travel destination has a maximum value (default to 250 if the information is not given). '
77        + 'The user can specify a minimum number of baths, beds or bedrooms (omit these search parameters if the information is not given). '
78        + 'Inform the user of the assumptions you made to make the Airbnb search.'
79        + 'Inform the user the datasetId and the amount of items you received from the airbnb_search tool in order to explore the results later. '
80        + "As your response, return only a JSON object (and nothing more! not even a ```json or ``` wrapper) with the keys 'datasetId' and 'assumptions', with their corresponding values in string format, along with the key 'totalItems' with the number of items in the dataset in number format."
81      ],
82      ['placeholder', '{chat_history}'],
83      ['human', '{input}'],
84      ['placeholder', '{agent_scratchpad}'],
85    ]);
86  }
87}
88
89export default ResearcherAgent;

1import { Log, ApifyClient } from 'apify';
2import { ChatPromptTemplate } from '@langchain/core/prompts';
3import { createToolCallingAgent, AgentExecutor } from 'langchain/agents';
4import { ChatOpenAI } from '@langchain/openai';
5import { StructuredToolInterface } from '@langchain/core/tools';
6import { CostHandler } from '../utils/cost_handler.js';
7
8/**
9 * Interface for parameters required by SuccessAgent class.
10 */
11export interface SuccessAgentParams {
12  apifyClient: ApifyClient,
13  modelName: string,
14  openaiApiKey: string,
15  log: Log | Console;
16}
17
18/**
19 * An AI Agent summarizes the efforts of all agents to answer the users's question.
20 */
21export class SuccessAgent {
22  protected log: Log | Console;
23  protected apifyClient: ApifyClient;
24  public agentExecutor: AgentExecutor;
25  public costHandler: CostHandler;
26
27  constructor(fields?: SuccessAgentParams) {
28    this.log = fields?.log ?? console;
29    this.apifyClient = fields?.apifyClient ?? new ApifyClient();
30    this.costHandler = new CostHandler(fields?.modelName ?? 'gpt-4o-mini', this.log);
31    const llm = new ChatOpenAI({
32      model: fields?.modelName,
33      apiKey: fields?.openaiApiKey,
34      temperature: 0,
35      callbacks: [
36        this.costHandler,
37      ],
38    });
39    const tools = this.buildTools();
40    const prompt = this.buildPrompt();
41    const agent = createToolCallingAgent({
42      llm,
43      tools,
44      prompt,
45    });
46    this.agentExecutor = new AgentExecutor({
47      agent,
48      tools,
49      verbose: false,
50      maxIterations: 3,
51    });
52  }
53
54  protected buildTools(): StructuredToolInterface[] {
55    // Tools are initialized to be passed to the agent
56    return [];
57  }
58
59  protected buildPrompt(): ChatPromptTemplate {
60    return ChatPromptTemplate.fromMessages([
61      ['system',
62        'You are a experienced travel agent that wants to help the user plan the perfect trip. '
63        + 'You asked some colleagues for help and they sent you a bunch of info to help the user.'
64        + 'The user may have specified other requests like pets, bedrooms, beds, baths, gym, etc so make sure to mention them. '
65        + 'Make sure that you show the rating, price, image, description, amenities (or other useful information) and a link to view more information about the chosen results. '
66        + 'If more than one location was specified in the search, try to make it so that your recommendations include results in every one of them. '
67        + 'Please explain why you chose those results and how many you explored before making your choice. '
68      ],
69      ['placeholder', '{chat_history}'],
70      ['human', '{input}'],
71      ['placeholder', '{agent_scratchpad}'],
72    ]);
73  }
74}
75
76export default SuccessAgent;