AskTable Private Deployment Complete Guide

If you're selecting an AI data analysis tool for your enterprise, you'll likely encounter the same question:

"Can our business data stay on our own servers?"

This question comes from compliance heads at financial companies, CTOs at healthcare enterprises, technical managers at government departments - they share a common request: data must stay in our own hands.

90% of AskTable enterprise users choose private deployment.

---

I. Why Private Deployment Is Needed

Enterprise data security concerns typically come from three aspects:

1. •Compliance requirements: Finance, healthcare, government and other industries have clear regulatory requirements for data storage and processing
2. •Business secrets: Customer information, transaction records, and operational metrics are core enterprise assets that cannot be leaked
3. •Self-control: Critical business systems must be in their own hands, not dependent on third parties

AskTable private deployment's core principle is only one: all data remains in the enterprise's own environment.

Specifically:

• •Enterprise data: Remains in enterprise's local database, not transmitted to any external system
• •Conversation history: All user-AI interaction content stored locally
• •Metadata: Table structures and field information stored locally
• •Chain of thought logs: AI analysis reasoning steps and all operation records are local

---

II. Two Private Deployment Solutions

AskTable's private deployment is divided into two solutions, with core difference being whether AI model runs on cloud or on-premises.

Solution A: Cloud Model

This is what most enterprises choose.

All data stored locally, AI inference calls cloud models (like Claude, GPT, Qwen, etc.), only inference requests sent (not saved), optional SDI desensitization encrypted transmission.

Solution A advantages:

• •No need to purchase GPU servers, lower cost
• •Directly use top-tier AI models, strongest analysis capability
• •All enterprise data still local, model only infers not saves
• •Supports SDI desensitization encrypted inference, sensitive information automatically desensitized before transmission

Solution B: On-Premises Model

Data doesn't leave the premises, inference doesn't either.

All data stored and processed locally, AI model also runs on local GPU servers.

Solution B characteristics:

• •Highest level data security, no data leaves enterprise network
• •Can run completely offline
• •Requires GPU servers, higher hardware investment
• •Enhanced version supports local large models through upgrade

---

III. Hardware Configuration Requirements

Solution A: Cloud Model

Solution A doesn't need GPU servers, just a regular server.

Large teams recommended to split PostgreSQL, Redis, Qdrant and other infrastructure to different servers.

Solution B: On-Premises Model

Solution B needs, in addition to application servers, a GPU server to run local AI models (starting model: Qwen 235B).

Application server configuration (same as Solution A):

GPU server configuration (running Qwen 235B model):

---

IV. Pricing and Licensing

AskTable private deployment offers multiple editions covering different team size needs, supporting annual subscription and permanent license.

For detailed edition, pricing and licensing information, please check the Pricing page.

---

V. Where to Deploy

AskTable private deployment supports two deployment locations:

1. Cloud Private Deployment

Deploy in your own cloud account, data and control all under your cloud account.

Recommended: Alibaba Cloud One-Click Deployment

Through Alibaba Cloud Market, you can quickly deploy a complete AskTable private environment:

• •Automatically configure infrastructure
• •Just enter AI model keys to use
• •Data remains under your Alibaba Cloud account

Alibaba Cloud Market One-Click Deployment

Alternative: Sealos One-Click Deployment

Using Sealos cloud operating system, you can also achieve quick deployment:

• •No cloud computing expertise needed
• •Complete deployment in 1 minute

Sealos One-Click Deployment

Cloud deployment recommendation: Suggest using cloud AI models (Solution A) directly, fully utilizing cloud network advantages while maintaining data security within your own cloud account.

2. On-Premises IDC Private Deployment

Deploy in your own data center or private cloud environment, suitable for enterprises with complete IT infrastructure.

• •Delivered through unified installation package, Docker Compose one-click startup
• •Includes all necessary components (frontend, backend, database, cache, vector database)
• •Supports sub-path deployment (embedded in existing systems)
• •Supports custom branding (app name, title)
• •Supports domestic network environment (built-in domestic mirror sources)

On-premises deployment recommendation: Also suggest using cloud AI models (Solution A), ensuring transmission security through SDI desensitization encryption. For particularly sensitive scenarios, consider upgrading to Solution B.

View On-Premises IDC Deployment Documentation

Deployment Method Comparison

---

VI. Integration with WeChat Work and Other Internal Systems

WeChat Work integration requires private deployment version only. SaaS version cannot integrate with WeChat Work.

Why Doesn't SaaS Support This?

WeChat Work integration requires deploying services within the enterprise's internal network to interface with WeChat Work APIs. This inherently requires a private deployment environment - SaaS version runs on cloud and cannot directly connect with enterprise internal networks.

Private Deployment Integration Capabilities

---

VII. Selection Suggestions

Quick decision based on your actual situation:

Step 1: Confirm data security requirements

• •Data must stay local → Private deployment
• •Can accept cloud storage → SaaS version

Step 2: Choose AI model solution

• •Want strongest AI capability while ensuring data doesn't leave premises → Solution A (cloud model, recommended)
• •Compliance requires inference process also doesn't leave premises → Solution B (on-premises model, needs enhanced version)

Step 3: Choose deployment location

• •No complete IT infrastructure → Cloud private deployment (Alibaba Cloud one-click recommended)
• •Have IDC data center and operations team → On-premises IDC deployment

Step 4: Choose edition

View Pricing page for detailed editions and pricing.

---

Related Reading:

• •AskTable Data Privacy Protection
• •Enterprise Data Security Best Practices
• •Deploy AskTable on Sealos in 1 Minute

Contact Us:

• •Website: asktable.com
• •Email: hi@datamini.ai