# Provably

> Provably makes agentic workflows reliable. SourceryKit, its Python SDK (`sourcerykit` on PyPI, import as `provably`), adds a deterministic eval layer to any Python agent. It monkey-patches outbound HTTP calls (requests, httpx, aiohttp), records every interaction, enforces a trusted endpoint registry, and lets receiving agents evaluate handoff claims against cryptographically proven query records — so multi-agent workflows stay hallucination-free without model self-evaluation.

- **SDK:** SourceryKit
- **Distribution:** `sourcerykit` (PyPI publish pending), import as `provably`
- **Python:** 3.11+
- **License:** Proprietary
- **Status:** v0.2

SourceryKit is built around five pillars: `init` (bootstrap and interceptor install), `intercept` (monkey-patched HTTP recording), `handoff` (typed claim payloads between agents), `eval` (deterministic guardrail checks against Provably query records), and `trusted_endpoints` (URL allowlist enforcement before requests leave the process).

Framework coverage includes OpenAI SDK, Anthropic SDK, Pydantic AI, LangChain, LangGraph, LlamaIndex, AutoGen, Haystack, Phidata/Agno, OpenAI Agents SDK, Google GenAI, LiteLLM, DSPy, smolagents, and CrewAI — all via transport-level patching of requests, httpx, and aiohttp.

Four verification modes: `verbatim` (canonical JSON equality), `field_extraction` (equality at a JSON path), `schema_type` (JSON schema validation), and `range_threshold` (numeric bounds check). Outcomes are `PASS`, `CAUGHT`, or `ERROR`.

## SDK

- [AGENTS.md](https://github.com/ProvablyAI/sourcerykit/blob/main/AGENTS.md): Agent onboarding instructions — also mirrored in full on the machine view at https://provably.ai/ai
- [GitHub Repository](https://github.com/ProvablyAI/sourcerykit): Source code, examples, tests, and full README
- [SDK Architecture](https://github.com/ProvablyAI/sourcerykit/blob/main/docs/architecture.md): Dependency rules and design decisions
- [Changelog](https://github.com/ProvablyAI/sourcerykit/blob/main/CHANGELOG.md): Version history and release notes
- [Context](https://github.com/ProvablyAI/sourcerykit/blob/main/CONTEXT.md): Project context document
- [Examples](https://github.com/ProvablyAI/sourcerykit/tree/main/examples): Working integration examples including OpenAI Agents

## Machine View

- [Machine view of the homepage](https://provably.ai/ai): Plain-text, markdown-styled mirror of provably.ai built for LLMs and agents — same content, no scripts or imagery required to read it

## Docs

- [Developer Documentation](https://provably.ai/docs): Introduction, architecture, and platform guides
- [API Reference (Swagger)](https://api.provably.ai/api/v1/swagger/): Full REST API documentation
- [Provably Architecture](https://provably.ai/docs/provably-architecture): System architecture and middleware design
- [Cryptography](https://provably.ai/docs/cryptography): Cryptographic foundations
- [Connecting Your Data Source](https://provably.ai/docs/connecting-your-data-source): Database setup guide
- [Querying Data](https://provably.ai/docs/querying-data): Running verifiable SQL queries
- [Verifying Proofs](https://provably.ai/docs/verifying-proofs): Proof verification guide
- [Prover & Verifier Performance](https://provably.ai/docs/prover_verifier_performance): Benchmarks

## Research

- [QEDB Paper (PDF)](https://provably.ai/papers/qedb-paper.pdf): Expressive and modular verifiable databases — compiles SQL directly over tabular data, proofs under 5kb, no SNARKs or circuits required

## Blog

- [Provably V2 is Live](https://provably.ai/blogs/V2-Live): V2 launch — relational DB support and verifiable SQL IDE
- [Beyond Circuits](https://provably.ai/blogs/Beyond_Circuits_New_Era_for_Verifiable_DBs): Why QEDB makes verifiable SQL practical
- [Rethinking Verifiable Databases](https://provably.ai/blogs/Rethinking_Verifiable_Databases): QEDB as a third path beyond ADS and SNARK-based systems

## Company

- [About Us](https://provably.ai/about-us): Team, mission, and background
- [Changelog](https://provably.ai/changelog): Product updates