Skip to main content

jsonbit

Rust TypeScript Python MCP License: MIT

jsonbit is an in-memory and on-disk analytical engine written in Rust that normalizes nested .json and .jsonl payloads into Third-Normal-Form (3NF) bit-packed columnar tables. It provides sub-10-microsecond bitmap query execution, compact .jsonbit binary snapshots, and a compact OLAP schema summary designed for Large Language Model (LLM) context windows and Model Context Protocol (MCP) agents.


Table of Contents


Overview

Passing large structured datasets (API responses, telemetry logs, database exports) directly into LLM prompts as raw JSON or text tables introduces two bottlenecks:

  1. Token Volume: A 2,000-record dataset consumes 50,000–180,000 tokens in JSON or text-table encodings, increasing API cost and time-to-first-token (TTFT).
  2. In-Context Arithmetic Errors: Autoregressive attention struggles with exact row counting, multi-column filtering, and floating-point summation across thousands of lines.

jsonbit addresses both problems by separating storage and computation from prompt context:

  • Automatic 3NF Normalization: Decomposes nested 1-to-1 objects, sparse optional fields, dynamic-keyed maps, and 1-to-Many child arrays into flat columnar tables linked by Compressed Sparse Row (CSR) offset arrays.
  • Bit-Packed Columnar Storage: Compresses integers using Frame-of-Reference (FOR) delta packing, strings using dictionary bit-packing with inverted 64-bit bitmaps (BitVec64), and booleans using 1 bit per row—reducing storage size by 64.4% compared to minified JSON while maintaining lossless round-trip reconstruction.
  • Compact OLAP Data Cards: Emits a ~200–540 token summary containing table schemas, column types, exact numeric aggregates (sum, avg, min, max), and categorical value distributions (99.24% fewer tokens than full-dataset text encodings).
  • Deterministic Execution: Evaluates natural-language slices and SQL queries (SELECT, WHERE, GROUP BY, ORDER BY, LIMIT, and parent-child JOINs) directly on bit-packed columns and 64-bit word-aligned bitmaps in 8.0 µs average latency.

Installation & Setup

Prerequisites

  • Rust (1.75+, rustup.rs)
  • (Optional) Python 3.10+ for the Python SDK, or Node.js 20+ for the TypeScript SDK

Build & Install Locally

git clone https://github.com/prathikmurugan/jsonbit.git
cd jsonbit

# Build optimized release binary (./target/release/jsonbit)
cargo build --release

# Install `jsonbit` globally into your PATH (~/.cargo/bin)
cargo install --path .

Verify installation:

jsonbit --help

(Optional) Install the Python package & binary in editable/wheel mode via maturin or pip:

pip install .

Agent Integrations

jsonbit includes a built-in stdio JSON-RPC 2.0 MCP server (jsonbit mcp) with zero external runtime dependencies. It exposes five tools: olap_load, olap_digest, olap_ask, olap_sql, and olap_rehydrate.

OpenCode (opencode.json)

Add the following to opencode.json in your project root (or globally in ~/.config/opencode/opencode.json):

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "jsonbit": {
      "type": "local",
      "command": [
        "/path/to/jsonbit/target/release/jsonbit",
        "mcp"
      ],
      "enabled": true
    }
  }
}

Claude Code (.mcp.json or CLI)

Register via the Claude Code CLI:

claude mcp add jsonbit -- /path/to/jsonbit/target/release/jsonbit mcp

Or include in .mcp.json:

{
  "mcpServers": {
    "jsonbit": {
      "command": "/path/to/jsonbit/target/release/jsonbit",
      "args": ["mcp"]
    }
  }
}

Cursor & Windsurf

Add the .mcp.json configuration above to .cursor/mcp.json or ~/.codeium/windsurf/mcp_config.json.

OpenAI Codex CLI (AGENTS.md)

Codex and terminal agents automatically read AGENTS.md in the repository root to invoke jsonbit via CLI or MCP when inspecting .json, .jsonl, or .jsonbit files.


SDK Usage

Python (python/jsonbit)

Requires Python 3.10+ with no third-party pip packages:

from jsonbit import JsonBit  # or `from python.jsonbit import JsonBit` from repo root

with JsonBit("benchmarks/data/github-repos.json") as olap:
    # Generate compact OLAP schema & statistical card
    card = olap.card()

    # Deterministic scalar answer
    ans = olap.answer("How many repositories have more than 150000 stars and more than 30000 forks?")
    assert ans == "31"

    # Read-only SQL query (supports GROUP BY, ORDER BY, LIMIT, and JSON dot-paths)
    rows = olap.sql(
        "SELECT defaultBranch, COUNT(*) AS cnt FROM records GROUP BY defaultBranch ORDER BY cnt DESC"
    )

    # Save compressed .jsonbit binary snapshot
    olap.pack("repos.jsonbit")

TypeScript / Node.js (@jsonbit/core)

import { JsonBit, rehydrateFrom3NF } from '@jsonbit/core'

const engine = JsonBit.from(jsonData)

// Retrieve the OLAP data card
console.log(engine.getDataCard().text)

// Retrieve a verified query slice or scalar answer
console.log(engine.answer('How many shipped orders include a USB Cable?'))
console.log(engine.slice('How many shipped orders include a USB Cable?').context)

// Execute SQL (automatically resolves JSON dot-paths such as customer.name -> customer_name)
const rows = engine.sql(`SELECT orderId, customer.name, total FROM orders WHERE total > 500`)

// Losslessly reconstruct original nested JSON
const original = rehydrateFrom3NF(engine)
engine.close()

Rust Library (jsonbit)

use jsonbit::{execute_sql, JsonBitEngine};
use serde_json::Value;

let data: Value = serde_json::from_str(raw_json)?;
let engine = JsonBitEngine::from_json(&data);

let card = engine.data_card();
let slice = engine.query_slice("How many active employees work in Sales?");
let rows = execute_sql(&engine, "SELECT department, COUNT(*) AS cnt FROM employees GROUP BY department")?;
let restored = engine.rehydrate();

CLI Reference

Command Arguments Description
jsonbit card <file.json|.jsonl|.jsonbit> Prints the $O(1)$ 3NF OLAP Cube Card (~250 tokens) and storage compression statistics.
jsonbit digest <file.json|.jsonl|.jsonbit> Prints the 1-Turn Factorized 3NF Holistic Digest (100% of rows + BitVec64 Pivot Legend, -41% to -72% smaller than TOON).
jsonbit ask <file> "<question>" Evaluates a question against the bitmap index and prints the verified context slice.
jsonbit answer <file> "<question>" Evaluates a question and prints only the scalar answer value.
jsonbit sql <file> "<SELECT ...>" Executes a read-only SQL query against the 3NF columnar store (WHERE, IN, BETWEEN, LIKE, GROUP BY, ORDER BY, LIMIT, OFFSET, CSR JOIN).
jsonbit pack <input.json> <output.jsonbit> Serializes the bit-packed 3NF tables into a binary .jsonbit snapshot.
jsonbit rehydrate <file.json|.jsonl|.jsonbit> Reconstructs and prints the original nested JSON document.
jsonbit mcp (none) Starts the concurrent stdio JSON-RPC 2.0 MCP server (olap_load, olap_digest, olap_ask, olap_sql, olap_rehydrate).

Architecture

flowchart TD
    Input["Nested JSON / JSONL Input"] --> Normalizer["3NF Normalizer & CSR Indexer (rust/normalizer.rs)"]
    Normalizer --> ColStore["Bit-Packed Columnar Store (rust/column.rs)"]
    ColStore --> Bitmap["BitVec64 Bitmap Engine (rust/bitmap.rs)"]
    Bitmap --> Card["OLAP Data Card (rust/engine.rs)"]
    Bitmap --> SqlExec["SQL & Predicate Executor (rust/query_exec.rs)"]
    ColStore --> Snapshot[".jsonbit Binary Snapshot (rust/snapshot.rs)"]
    ColStore --> Rehydrate["Lossless 3NF Rehydrator (rust/normalizer.rs)"]

Storage Layout (rust/column.rs & rust/normalizer.rs)

  1. ForIntColumn (Frame-of-Reference Integers): Computes [min_val, max_val] per column and packs unsigned deltas val - min_val into ceil(log2(max_val - min_val + 1)) bits per row inside a Vec<u64> buffer. Column headers store sum, min, max, and non_null_count for O(1) unfiltered aggregations.
  2. DictStringColumn (Dictionary + Inverted Bitmaps): Interns unique strings into a string arena, bit-packs dictionary IDs into ceil(log2(|D|)) bits per row, and builds a 64-bit word-aligned BitVec64 inverted index per category (|D| <= 64) for O(1) equality lookups and POPCNT frequency histograms.
  3. BoolBitmapColumn (1-Bit Booleans): Packs 64 boolean values into each u64 word alongside a 1-bit validity mask.
  4. CSR 1-to-Many Join Index: Child arrays (e.g., orders[].items[]) are normalized into child tables with a contiguous csr_offsets: Vec<u32> array mapping each parent row i to its child slice csr_offsets[i]..csr_offsets[i + 1].

Benchmarks

All benchmarks execute directly against the compiled Pure-Rust jsonbit binary (target/release/jsonbit).

1. Live gemini-3.8-flash Evaluation Across 10 Real-World Workloads (27 Questions)

Evaluates Raw JSON, Minified JSON, YAML, TOON (@toon-format/toon), and jsonbit (Pure-Rust olap_load + olap_sql) across 10 production workloads (W1–W10), including flat tables, 1-to-Many e-commerce orders, sparse .jsonl OpenTelemetry logs with free-text messages, singleton Kubernetes configs, IAM primitive arrays, wide 30-column tables, and multi-byte UTF-8 financial ledgers:

Live Gemini 3.8 Flash Accuracy and Token Efficiency

Format / Tool Mode Live gemini-3.8-flash Accuracy Total Prompt Tokens Token Reduction vs. Raw JSON Token Reduction vs. TOON
Raw JSON (Pretty) 1-Turn Full Text 55.6% (15/27) 375,713 — —
Minified JSON 1-Turn Full Text 48.1% (13/27) 237,891 -36.7% —
YAML 1-Turn Full Text 59.3% (16/27) 293,690 -21.8% —
TOON v4 (@toon-format/toon) 1-Turn Full Text 59.3% (16/27) 169,627 -54.9% —
jsonbit Holistic Digest (olap_digest) 1-Turn (100% Rows + Pivot Legend) 96.3% (26/27) 101,302 -73.0% -40.3%
jsonbit Cube Card + SQL (olap_load + olap_sql) Agentic $O(1)$ Card + Rust SQL 100.0% (27/27) 11,729 -96.9% -93.1%

Qualitative & Holistic Executive Summary Verification (1-Turn: TOON vs. jsonbit digest)

For qualitative tasks where the LLM must read all records in a single turn and write a multi-fact executive summary (naming all departments, products, microservice failure messages, and international counterparties across W3, W4, W5, and W10):

Holistic Synthesis Task TOON v4 Facts Preserved jsonbit digest Facts Preserved TOON Tokens jsonbit digest Tokens Token Savings vs. TOON
H1 — HR Directory Executive Summary 8 / 9 9 / 9 (100%) 4,881 4,372 -10.4%
H2 — E-Commerce Catalog & Orders Summary 8 / 8 8 / 8 (100%) 15,405 6,933 -55.0%
H3 — Production Incident Log Triage Report 9 / 9 9 / 9 (100%) 13,665 4,169 -69.5%
H4 — International Settlement Ledger Summary 2 / 6 6 / 6 (100%) 3,781 2,403 -36.4%
Total Across All Qualitative Summaries 27 / 32 (84.4%) 32 / 32 (100.0%) 37,732 17,877 -52.6%

2. Token Scaling Curve (4 $\rightarrow$ 25,000 Nested Orders)

While linear text encodings (Raw JSON, Minified JSON, YAML, TOON) scale at $O(N)$ with row count and exceed LLM context windows at scale, jsonbit maintains an $O(1)$ prompt footprint (~67–473 tokens):

LLM Context Tokens vs Dataset Scale

Dataset Scale (Nested Orders) Raw JSON (Tokens) YAML (Tokens) TOON v4 (Tokens) jsonbit Rust Card (Tokens) Rust Ingest Time
4 rows (Tiny API Payload) 166 122 71 67 0.04 ms
50 rows (~150 line items) 7,780 5,801 4,783 447 0.52 ms
500 rows (~1,500 line items) 78,115 58,244 47,983 449 4.72 ms
5,000 rows (~15,000 line items) 783,443 366,356 293,972 472 58.97 ms
25,000 rows (~75,000 line items) 3,917,366 1,831,843 1,469,911 473 334.42 ms

3. Memory Footprint & Query Speed (Pure-Rust jsonbit vs. jq CLI)

Pure-Rust Memory Footprint and Query Speed

Workload Category Minified JSON Rust 3NF Bit-Packed Pretty JSON (Tok) TOON (100% Rows) jsonbit digest (100% Rows) jsonbit card ($O(1)$) 3NF Lossless
W1-tiny-api-response Tiny Payload (<500B) 0.3 KB 0.1 KB 166 71 67 (-5.6%) 67 PASS
W2-singleton-k8s-config Singleton Config Tree 0.3 KB 0.2 KB 167 131 110 (-16.0%) 110 PASS
W3-hr-payroll-directory Tabular (200 rows) 28.4 KB 4.5 KB 14,064 4,847 4,145 (-14.5%) 272 PASS
W4-ecommerce-orders-1toN 1-to-Many (150 orders) 48.3 KB 4.5 KB 25,466 15,372 6,878 (-55.3%) 472 PASS
W5-otel-production-logs Sparse .jsonl + Free-Text 44.4 KB 4.7 KB 17,439 13,632 4,123 (-69.8%) 226 PASS
W6-github-repos-metrics High-Card + Ranges 28.6 KB 3.4 KB 14,829 5,364 4,285 (-20.1%) 309 PASS
W7-stripe-dispute-webhooks Financial Webhooks 28.9 KB 3.1 KB 14,333 4,104 3,906 (-4.8%) 250 PASS
W8-iam-primitive-arrays Primitive Arrays 14.4 KB 1.6 KB 7,185 4,810 1,247 (-74.1%) 164 PASS
W9-wide-enterprise-crm Wide Table (30 cols) 68.6 KB 5.7 KB 40,779 9,806 9,516 (-3.0%) 1,136 PASS
W10-unicode-financial-ledger UTF-8, Nulls & Zeroes 18.4 KB 2.9 KB 9,550 3,750 2,359 (-37.1%) 243 PASS
Total (10 Workloads) All Categories 280.6 KB 30.6 KB (-89.1%) 143,978 61,887 36,636 (-40.8%) 3,249 (-94.7%) 10/10 PASS
  • Query Latency (Pure-Rust BitVec64 SQL vs. jq C CLI):
    • jq CLI (stedolan/jq): 6,545.3 µs average per query
    • jsonbit Pure-Rust SQL (execute_sql): 6.8 µs average per query (963x faster)
    • jsonbit Pure-Rust Bitmap Slice (query_slice): 19.9 µs average per query (329x faster)
  • Standard 8-Dataset Benchmark Suite:
    • Storage Compression: 1,178,283 bytes (Minified JSON) $\rightarrow$ 419,661 bytes (-64.4% RAM/storage reduction), 8/8 (100%) lossless 3NF rehydration.
    • Token Reduction: 334,046 tokens (TOON v4) $\rightarrow$ 2,536 tokens (-99.24%).
    • 239-Question Evaluation Accuracy: 99.16% (237/239) strict match at 8.0 µs average query latency.

4. Autonomous Coding Agent Head-to-Head (OpenCode v1.19.5)

Evaluated in isolated directories with zero prompt hints and zero AGENTS.md/CLAUDE.md files (opencode/big-pickle) across both standard (bash-enabled) and sandboxed (bash: "deny") agent environments:

Agent Mode jsonbit MCP (opencode.json only) TOON v4 (toon_encode / CLI) Raw JSON
Standard Agent (6 Scenarios, 18 Runs) 6/6 (88.1s) · 0 scripts 6/6 (142.8s) · awk/Python parsers 6/6 (93.7s) · Python scripts
Sandboxed Read-Only (bash: "deny", 8 Scenarios) 8/8 (77.7s) 7/8 (333.8s) · failed 500-order sum 7/8 (345.4s) · failed 500-order sum

License

MIT License.

Metadata

Release files for jsonbit 1.1.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for jsonbit 1.1.1
File Size Uploaded Provenance Actions
jsonbit-1.1.1.tar.gz 192.9 kB With provenance Details for jsonbit-1.1.1.tar.gz

Built distributions (wheels)

Table of built distributions (wheels) for jsonbit 1.1.1
File Actions
jsonbit-1.1.1-py3-none-win_amd64.whl Python 3 none Windows x86-64 Details for jsonbit-1.1.1-py3-none-win_amd64.whl
jsonbit-1.1.1-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl Python 3 none Linux glibc 2.17+ x86-64 Details for jsonbit-1.1.1-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl
jsonbit-1.1.1-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl Python 3 none Linux glibc 2.17+ ARM64 Details for jsonbit-1.1.1-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl
jsonbit-1.1.1-py3-none-macosx_11_0_arm64.whl Python 3 none macOS 11.0+ ARM64 Details for jsonbit-1.1.1-py3-none-macosx_11_0_arm64.whl
jsonbit-1.1.1-py3-none-macosx_10_12_x86_64.whl Python 3 none macOS 10.12+ x86-64 Details for jsonbit-1.1.1-py3-none-macosx_10_12_x86_64.whl

Total release size: 2.7 MB

Release files / jsonbit-1.1.1.tar.gz

Download URL jsonbit-1.1.1.tar.gz
Size 192.9 kB
Tags Source
SHA-256 checksum
How to use checksums
f7e742d5de2d634ce0148bbd761f6006cd4338919b2cf9adbd4da4620aceb2cb
BLAKE2b-256 checksum
How to use checksums
5d97481be6220be3c61206e408ed6490ba2eab5eb7efbd74cac221d970964d21
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 20, 2026.

Transparency log

Release files / jsonbit-1.1.1-py3-none-win_amd64.whl

Download URL jsonbit-1.1.1-py3-none-win_amd64.whl
Size 456.3 kB
Tags Python 3 Windows x86-64
SHA-256 checksum
How to use checksums
044cbbc9d1763df11fa4841a92cd3eab6035c7e345de8ccbbc877a3de1ffc563
BLAKE2b-256 checksum
How to use checksums
b49c8bb1642e691ef4f3b0e58a57499c59ee722529177478cee3223eda5c465e
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 20, 2026.

Transparency log

Release files / jsonbit-1.1.1-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl

Download URL jsonbit-1.1.1-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl
Size 537.7 kB
Tags Linux glibc 2.17+ x86-64 Python 3
SHA-256 checksum
How to use checksums
574752955f4f238aca84417df22c987624139366db5013795826e0a36068d282
BLAKE2b-256 checksum
How to use checksums
2832e535692765282a75310a5b1f7e73c3fa4caee88630622fe5dbc6c91cd77d
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 20, 2026.

Transparency log

Release files / jsonbit-1.1.1-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl

Download URL jsonbit-1.1.1-py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl
Size 497.9 kB
Tags Linux glibc 2.17+ ARM64 Python 3
SHA-256 checksum
How to use checksums
339e16fbef0511250c00b1ebdcc5947ddc2bf2a5f1c4564ec2b6c16f92ec21dd
BLAKE2b-256 checksum
How to use checksums
fae0913b47baf92f897c3a5afd9d314e36b844565c04b1cbe29c4418ee1da79c
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 20, 2026.

Transparency log

Release files / jsonbit-1.1.1-py3-none-macosx_11_0_arm64.whl

Download URL jsonbit-1.1.1-py3-none-macosx_11_0_arm64.whl
Size 494.3 kB
Tags Python 3 macOS 11.0+ ARM64
SHA-256 checksum
How to use checksums
e36c9f61b38855dddba08635494dbabec4c590d0c0eaed8ef12cac536a3ba69b
BLAKE2b-256 checksum
How to use checksums
0bef4b64d1d615160491b6665fc94ea91b09dd5e463e6de17ff9d51c91d31973
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 20, 2026.

Transparency log

Release files / jsonbit-1.1.1-py3-none-macosx_10_12_x86_64.whl

Download URL jsonbit-1.1.1-py3-none-macosx_10_12_x86_64.whl
Size 528.4 kB
Tags Python 3 macOS 10.12+ x86-64
SHA-256 checksum
How to use checksums
6bb360e70d17bf1b62e18024915d714611b58e7d92f7193e68be47e8d4beda98
BLAKE2b-256 checksum
How to use checksums
919066965c08aad00cd15182859e32adfa0c06af865142a0145fcfb232d225de
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 20, 2026.

Transparency log

Security for jsonbit 1.1.1

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

All 6 files in this release have been uploaded with provenance metadata.

Select from 6 files

Attestation for jsonbit-1.1.1.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 20, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page