AI memory isn't enough. SRT-1 gives you cryptographic provenance over your intellectual contributions. We index your codebase, tag risk, and build a cryptographically signed Trust Chain — proving your team's ownership over everything your AI generates.
pip install srt1-core
One command. 7 pipeline stages. A signed manifest that proves your AI was working from ground truth.
One command. One signed manifest. Every AI tool aligned.
Parses Python, JavaScript, TypeScript, Go, Rust, Java, C#, and C/C++. Extracts every class, function, and cross-file dependency chain.
Every symbol gets an architectural role (ORCHESTRATOR, AUTH_SECURITY, SERVICE_LAYER…) and risk tags (AUTH_SENSITIVE, WRITES_TO_DB, DYNAMIC_EXECUTION…).
SHA-256 chained hashes across all 7 pipeline stages. Verify the manifest was produced by a clean, uninterrupted run — cryptographically provable.
Every artifact carries a cryptographic SCIA provenance signature. Detect tampering before trusting any AI output.
Pure Python stdlib. No pip install chain to audit. Ships as a single command: srt1-index --repo_path .
Auto-generates CLAUDE.md, .cursorrules, AGENTS.md, and copilot-instructions.md so every AI tool reads the same ground truth.
Track every task as a seed through its full lifecycle — planted → germinating → growing → bloomed. Never lose a task mid-session.
Native Model Context Protocol server for Claude. Every tool call is context-aware, grounded in your actual codebase manifest.
Detects duplicate functions, stale files, and missing docstrings. Auto-proposes governed remediations with full audit trail.
Install SRT-1 as a native app on your phone or tablet. Monitor your codebase, check seed status, and review coherence — from anywhere. No app store required.
Run the Command Center in your browser or install it as a desktop app. Persistent access to your codebase intelligence without keeping a terminal open.
SRT-1 runs entirely on your own system. No proprietary cloud backends. No VC-funded tracking. The AST indexing and SQLite database live exclusively on your machine.
Zero IP Leak. Your codebase never leaves your local environment. SRT-1 maps your architecture and protects your intellectual property without ever phoning home.
No middleware API skimming or hidden token taxes. You use your own LLM API keys directly. You control the cost, the context, and the connection.
Full AST parsing for Python. Regex + structure extraction for everything else.
Run the middleware and open localhost:7605/dashboard. The Command Center shows real-time data from your repo — file counts, symbol inventory, seed farm, coherence score — updating as you work.
You're on the train with an idea. Open the PWA, type "add dark mode to the settings page," and plant the seed. SRT-1 generates a full-context blueprint from your codebase — then you review and approve it before anything is dispatched. Once approved, the Execution Bridge sends it to whatever code assistant is already running — Cursor, Claude, Copilot. You check back later: 🌸 BLOOMED — 3 files modified.
Session continuity built in. Every session ends with a handoff summary — what you accomplished, what's pending, and the exact prompt to restart. Close a conversation, come back a week later, paste one line and your AI picks up instantly. No rethinking, no re-explaining.
from srt1_code_indexer import SRT1CodeIndexer indexer = SRT1CodeIndexer(repo_path=".") manifest = indexer.index_repository() # Every symbol, tagged and signed print(manifest["metadata"]["total_symbols_indexed"]) # 661 print(manifest["metadata"]["pipeline_stages"]) # 7 print(manifest["trust_chain"][-1]["stage"]) # "manifest"
The full engine runs free on your machine. Upgrade when you need multi-module orchestration or team sync.
SRT-1 is built to protect the developer — the Original Seed — and their creations. The engine runs entirely local. Anonymous telemetry is opt-in only, and every payload is logged to your dashboard in real-time. If the engine ever transmits anything beyond what's disclosed, it flags itself as a violation.
Upload a ChatGPT JSON export. Paste a Claude conversation. Get every abandoned idea back, prioritized and ready to pick up — no setup required.
Try Conversation Recovery →