goalfydata

A shared place for AI agents to build, update, analyze, and reuse business data.

Turn spreadsheets, APIs, databases, and agent outputs into reusable datasets and data apps

that preserve business context and stay up to date.

Website

·

Get Started

·

Documentation

---

Understand GoalfyData in 30 Seconds

Codex, Claude Code, Manus, and other connected agents can create datasets, write update scripts, analyze results, and build data apps. GoalfyData keeps the resulting data together with its field definitions, metric definitions, table relationships, permissions, and governance rules.

The result is a durable data asset that can be reused across conversations, agents, devices, and teams. Import data, run SQL analysis, schedule updates, share with controlled access, and deploy data apps from the same dataset. When the dataset updates, connected apps continue to read the latest data.

Quick Start

The fastest path is to open the integration page for your platform, copy its setup instructions, and give them to your agent. Create your API Key in [GoalfyData Settings](https://goalfydata.ai/settings); keys use the gfk_ prefix and are shown only once.

| Platform | Fastest setup | Detailed guide | Status |

|---|---|---|---|

| Codex | [Open the Codex integration](https://goalfydata.ai/integrations/codex) and send the setup text to Codex | [Codex Quick Start](./docs/codex-quickstart.md) | Available |

| Claude Code | [Open the Claude Code integration](https://goalfydata.ai/integrations/claude-code) and send the setup text to Claude Code | [Claude Code Quick Start](./docs/claude-code-quickstart.md) | Available |

| Manus | [Open the Manus integration](https://goalfydata.ai/integrations/manus), then add the MCP connector and upload the Skill in Manus | [Manus Quick Start](./docs/manus-quickstart.md) | Available |

| Other Agents / Generic MCP | Connect the remote MCP and load the generic Skill | [Generic Integration Guide](./generic/README.md) | Available for compatible MCP/CLI agents |

> Manus setup currently requires manual steps in its web interface; it cannot be completed by pasting an install runbook into a Manus conversation.

Minimal manual CLI setup

For macOS or Linux developers who prefer manual setup:

curl -fsSL https://cdn.goalfydata.ai/dataset-uds/install.sh | sh
uds-cli login --api-key gfk_your_api_key --api-url https://api.goalfydata.ai

Then follow the platform guide above to install the Skill/plugin and connect MCP. The detailed guides cover Windows, updates, key rotation, and troubleshooting. A successful connection exposes 20 GoalfyData MCP tools, including uds_query and uds_dataset_manage.

How AI Agents Create and Reuse Data Assets with GoalfyData

CSV files give an agent rows and columns. Prompts give it instructions for one conversation. Neither preserves the structure, definitions, relationships, and business rules another agent needs to continue the work later. GoalfyData turns that missing context into a reusable data asset.

1. Connect GoalfyData with your AI agent

Choose your platform in [Quick Start](#quick-start), copy the MCP and Skill setup instructions, and give them to Codex, Claude Code, Manus, or another compatible agent.

This connection gives your agent the ability to create, understand, update, manage, and reuse governed data assets—not simply send data to another tool.

2. Tell your agent what you want to build

Start with the business outcome and data you need. For example:

> Help me scrape today's recommended AI tools and categories on GitHub.

Your agent can work out the required sources, tables, field relationships, and metrics, then collect or prepare the data. You can also provide an Excel or CSV file, or connect sources such as Google Analytics 4 (GA4) and Google Search Console (GSC).

Want to try the workflow yourself? Click the file title below to download the example files, then give them to your agent and describe the data asset you want to build.

• [EXAMPLE Ecommerce Data](https://github.com/GoalfyAI/goalfydata/raw/refs/heads/main/examples/sample-data/example-ecommerce-data.zip) — simulated Amazon, Shopify, Meta Ads, returns, and SKU cost data for two weekly reporting periods

For other ready-to-use datasets, email [goalfydata@goalfyai.com](mailto:goalfydata@goalfyai.com).

3. Create a reusable data asset

Ask your agent to organize the result in GoalfyData. It saves more than rows of data: it preserves the context that another agent needs to use the asset correctly.

| Preserved context | What it provides |

|---|---|

| Tables | The original data structure |

| Relationships | How tables and records connect |

| Field meanings | What each field represents |

| Rules | Business logic, metrics, and processing requirements |

| Usage guidance / Skills | How agents should query, update, and apply the asset |

Your AI agent doesn't just access data. It understands how to use it.

4. Create applications powered by your data assets

Once the data asset exists, ask your agent to create a purpose-built application such as an analysis report, business dashboard, or automated analysis tool.

For example:

> Create a dashboard from the GitHub AI Tools Intelligence dataset.

The agent uses the asset's structure, field definitions, relationships, and analysis rules to build the application. The application remains connected to its dataset, so when the data updates, metrics recalculate and reports or dashboards continue to reflect the latest information—without rebuilding the application or manually changing its configuration.

You can [view the dashboard from this walkthrough](https://app-08a1b21d0a98-github-ai-tools-dashboar-5f51cc722fed.goalfydata.app/).

To keep the data current, tell your agent the update schedule in natural language. The agent will explain Managed Refresh usage and ask for confirmation before enabling it.

5. Analyze, manage,