Question 1

What exactly is Savant Chat?

Accepted Answer

Savant Chat is an AI-powered co-pilot for smart-contract security. It reviews smart contract code in real time, flags exploits before they hit mainnet, and stores audit proofs on-chain. Using a sophisticated multi-agent AI system, Savant Chat analyzes code across 200+ vulnerability classes—the most comprehensive coverage in the industry.

Question 2

How do I get started?

Accepted Answer

1) Sign up at savant.chat and receive $75 in free credits (no credit card required), 2) Upload your smart contract code as .zip, .sol files, or connect your GitHub repository, 3) Select the audit scope and click "Audit", 4) Receive a detailed security report with actionable recommendations in minutes.

Question 3

Is there a free tier?

Accepted Answer

Yes. Every new account gets $75 USD in free credits, enough to audit a medium-sized smart contract from start to finish. No credit card required.

Question 4

How does it work?

Accepted Answer

Upload your project, select the audit scope, and let our AI vulnerability scanner go to work. Our multi-agent system deploys thousands of specialized AI agents in parallel, each focusing on specific vulnerability classes. In minutes, you'll receive a detailed security report across 200+ vulnerability classes.

Question 5

How much does Savant Chat cost?

Accepted Answer

Average human audit costs around $20 per line of code. Savant Chat pricing varies from $0.07 to $0.5 per line of code. New users get $75 in free credits.

Question 6

How does it compare to manual audit costs?

Accepted Answer

Manual audits: $20K-$100K+, take 2-4 weeks. Savant Chat: ~80% less cost, results in minutes. Best practice: use both together.

Question 7

Do you accept crypto payments?

Accepted Answer

Yes! We accept stablecoins (USDC) and traditional payment methods.

Question 8

Do you offer enterprise pricing?

Accepted Answer

Yes. Custom enterprise plans with volume pricing, dedicated support, priority processing, custom integrations, and SLA guarantees. Contact us for details.

Question 9

What programming languages do you support?

Accepted Answer

Solidity (Ethereum, BSC, Polygon, all EVM chains), Vyper (DeFi protocols), and Rust (Solana, Near). Same 200+ vulnerability coverage across all languages.

Question 10

How can I integrate into my development workflow?

Accepted Answer

Add our GitHub Action or GitLab CI step into your pipeline. Savant Chat will scan each pull request diff and label issues by severity. Also available via API.

Question 11

Can I use it in CI/CD?

Accepted Answer

Yes. Seamless integration through GitHub Actions, GitLab CI, and API access. Every commit can be audited across 200+ vulnerability classes automatically.

Question 12

How long does an audit take?

Accepted Answer

It depends on the specific request and project scale — ranging from minutes to a couple of hours for larger projects.

Question 13

How is Savant Chat different from general AI models?

Accepted Answer

General AI models lack specific smart contract security context. Savant Chat uses a specialized multi-agent architecture with thousands of AI agents trained specifically for vulnerability detection across 200+ security classes.

Question 14

What makes you different from free tools like Slither?

Accepted Answer

Free tools: 15-20 patterns, high false positives, pattern-matching only. Savant Chat: 200+ vulnerability classes, semantic understanding, low false positives. Use free tools for first pass, Savant Chat for comprehensive coverage.

Question 15

What is the 200+ vulnerability class coverage?

Accepted Answer

We analyze across 200+ distinct vulnerability classes—the most comprehensive in the industry. Includes common ones (reentrancy, overflow) plus 180+ additional classes like oracle manipulation, flash loans, MEV vulnerabilities, governance exploits, etc.

Question 16

How do you compare to manual audits?

Accepted Answer

Speed: Minutes vs 2-4 weeks. Cost: ~80% less. Best practice: use both. AI for exhaustive detection, humans for complex business logic.

Question 17

Can I rely on Savant Chat alone without human audit?

Accepted Answer

No. Think of it as a "first-pass reviewer." AI finds edge cases; humans understand context. Strongest security = AI + human together.

Question 18

How secure is my data?

Accepted Answer

We transform and chunk your data before processing. We only work with AI providers that do NOT train on your data. All transmissions encrypted. Enterprise-grade security practices.

Question 19

Is my code stored or shared?

Accepted Answer

No. Your code is processed securely and not stored or shared. We work with trusted AI providers under strict agreements. After processing, your code is not retained.

Question 20

Can you audit private/proprietary code?

Accepted Answer

Yes. All code remains private and confidential. Enterprise plans available for additional security guarantees and custom SLAs.

Question 21

How accurate is Savant Chat?

Accepted Answer

Proven detection rates on independent CTFBench benchmarks. Outperforms other AI auditors and traditional tools. 200+ vulnerability coverage (vs 15-20 for free tools). Significantly lower false positives.

Question 22

What's in the audit report?

Accepted Answer

Severity ratings, detailed explanations, affected code locations, recommended fixes, confidence scores, vulnerability class classification, attack scenarios, estimated impact.

Question 23

Where can I get support?

Accepted Answer

Join our Discord community at discord.gg/pHfxVh9WSc for support and to connect with other builders.

Question 24

Do you have a community?

Accepted Answer

Join our Discord community at discord.gg/pHfxVh9WSc for support and to connect with other builders.

Question 25

How can I invite friends?

Accepted Answer

Go to "Referrals" in settings and share your unique link. You'll both receive bonuses.

Frequently Asked Questions

📋 Getting Started

What exactly is Savant Chat?

How do I get started?

Is there a free tier?

How does it work?

Have more questions?

Ready to Secure Your Smart Contracts?