Shayntech Excel AI Pro

AI access

Choose how to pay for AI

The trial is once per account. Each submitted task includes up to 41 AI calls over two hours. Started tasks count even if cancelled or the AI provider fails. Your own-key requests do not use your trial or plan allowance; your provider may charge you.

My default provider (own-key mode)

Model routing

Choose a provider and model by task

Auto uses your instruction: coding/reasoning first, then formulas, then bulk/data tasks, otherwise general. Use the chat selector to override it. One provider handles the whole task, including planning and repairs. Missing keys or provider errors stop the task; no automatic provider fallback. Bulk routing keeps the existing workbook read limits.

Your own AI provider

OpenRouter

Gemini

Groq

Claude

Qwen

Ollama

DeepSeek / OpenAI-compatible

Provider API base URL Model ID (must support tool calling) Your API key

Keys are saved on this device and sent directly to the selected provider. Use its official HTTPS API URL. The provider must allow browser requests and support Chat Completions; not every platform or model is compatible.

OpenRouter

API Key

Get free key at openrouter.ai → free models: DeepSeek R1, Llama 3.3, Gemma 3

Model

Google Gemini

API Key

Get free key at aistudio.google.com → Get API Key (generous free tier)

Model

Enter the exact active model ID from Google AI Studio. Agent tasks require function calling support.

Groq (Ultra-Fast)

API Key

Free key at console.groq.com — extremely fast inference

Model

Use a model available to your Groq account. Availability, context limits and pricing depend on the provider.

Anthropic Claude

API Key

Get key at console.anthropic.com — best reasoning and instruction following

Model

Qwen / Alibaba

API Key

Get from bailian.console.aliyun.com or your MaaS workspace

API Host (optional)

Leave empty for standard DashScope. Paste your openAiCompatible URL for MaaS endpoints (without /chat/completions)

Model

Ollama (Local / Cloud)

Ollama Base URL

For local PC: http://localhost:11434 | For cloud/remote server: use your server IP

Model Name

Run: ollama pull llama3.2 | Other options: mistral, codellama, qwen2.5, phi3

API Key (optional)

Only needed if your Ollama server is behind auth (e.g. ngrok with auth)

Agent Behaviour

Max agent loop iterations

How many actions the agent can take in one run before stopping

Context rows limit

Limit rows sent to AI to avoid token overflow

Require plan approval before execution

Agent shows its plan and waits for your ✅ before running any tools

🎨 Theme

Dark

Light

Midnight

Own-key tests send two small requests with synthetic data to each configured model. Provider charges may apply. Tests do not read or change your workbook.

Shayntech

AI Agent for Excel · v2.2.0

A fully agentic AI co-pilot that reads your workbook, understands your data, and autonomously takes action — all from a single chat command.

Capabilities

📖Reads bounded workbook snapshots and specific ranges, with values and formulas

✍️Writes values, sets Excel formulas, pastes data from files

🎨Formats cells — bold, colors, borders, number formats

📊Creates charts (bar, line, column, pie) from data ranges

📋Adds, renames, and manages worksheets

🔃Sorts and filters data ranges

📎Reads attached PDF, CSV, XLSX, TXT, JSON files

🤖Multi-step agent loop — plans, executes, verifies, reports

🔌OpenRouter · Gemini · Groq · Claude · Qwen/Alibaba · Ollama (local)

Quick Start

1. Go to Settings → pick your AI provider → enter API key → Save
2. Click "Test configured models" to check sample data, a formula and tool conversation
3. Come back to Chat → type your command
4. Watch the agent work step by step

Example commands:
• "Summarize the data in Sheet1 and add a total row"
• "Create a bar chart of column A vs column B"
• "Format all cells in row 1 as bold with green background"
• "Read the attached CSV and paste it into Sheet2 starting at A1"
• "Add a new sheet called Summary and write key metrics there"

Built by Shayntech · Your API key is stored locally only