Drift

A codebase intelligence platform that detects patterns, conventions, and violations in code while maintaining memory across sessions. Exposes an MCP server with six entry points (drift_scan, drift_status, drift_tool, drift_discover, drift_workflow, drift_explain) that give AI agents access to ~70 specialized analysis utilities covering violations, security patterns, call graphs, architectural boundaries, coupling metrics, taint flows, and more. Includes a Cortex memory system for semantic memory, knowledge retrieval, and multi-agent coordination. Built with Rust (NAPI-RS bindings) and TypeScript for performance.

Evaluated Mar 06, 2026 (0d ago) vunknown
Homepage ↗ Repo ↗ Developer Tools mcp code-analysis pattern-detection convention-enforcement security-audit coupling-analysis taint-analysis rust typescript codebase-memory
⚙ Agent Friendliness
60
/ 100
Can an agent use this?
🔒 Security
63
/ 100
Is it safe for agents?
⚡ Reliability
54
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
72
Documentation
55
Error Messages
50
Auth Simplicity
73
Rate Limits
50

🔒 Security

TLS Enforcement
80
Auth Strength
65
Scope Granularity
50
Dep. Hygiene
60
Secret Handling
58

Community/specialized tool. Apply standard security practices for category. Review documentation for specific security requirements.

⚡ Reliability

Uptime/SLA
60
Version Stability
55
Breaking Changes
50
Error Recovery
50
AF Security Reliability

Best When

You have a medium-to-large codebase and want AI agents to understand deep structural patterns, enforce conventions, and maintain persistent codebase intelligence across sessions.

Avoid When

Your codebase is small or in a language not yet supported. The v2 rewrite is under active development, so expect some rough edges.

Use Cases

  • Giving AI coding agents deep codebase awareness beyond file contents
  • Detecting pattern violations and enforcing coding conventions via AI assistants
  • Security auditing codebases through OWASP findings and taint flow analysis
  • Understanding module coupling, API contracts, and architectural boundaries
  • Maintaining persistent codebase knowledge across multiple AI agent sessions
  • Monte Carlo simulation for change impact estimation

Not For

  • Replacing a full CI/CD linting pipeline (complementary, not a replacement)
  • Non-code analysis tasks
  • Very small or trivial codebases where the overhead is not justified
  • Languages not yet supported (currently Python, Java, PHP, TypeScript, C#)

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

OAuth: No Scopes: No

No authentication required. Runs locally against your codebase. Supabase integration present in repo suggests optional cloud sync but not documented.

Pricing

Model: open_source
Free tier: Yes
Requires CC: No

No license file found in repository. Appears open source based on public GitHub repo. License terms unclear.

Agent Metadata

Pagination
unknown
Idempotent
Partial
Retry Guidance
Not documented

Known Gotchas

  • V2 is a significant rewrite - v1 is deprecated, v2 is under active development
  • README is minimal - real documentation is buried in drift v2/docs/
  • License terms are unclear (no LICENSE file found)
  • Supabase integration suggests potential cloud dependency not well documented
  • ~70 tools via drift_tool may overwhelm agent context - use drift_discover to filter
  • Requires initial drift_scan before other tools produce useful results
  • Docker and stdio transport modes available but HTTP setup requires Docker

Alternatives

Full Evaluation Report

Detailed scoring breakdown, competitive positioning, security analysis, and improvement recommendations for Drift.

$99

Scores are editorial opinions as of 2026-03-06.

5208
Packages Evaluated
26151
Need Evaluation
173
Need Re-evaluation
Community Powered