Codex

OpenAI Codex is an artificial intelligence model developed by OpenAI. It parses natural language and generates code in response. It is used to power GitHub Copilot, a programming autocompletion tool. Codex is a descendant of OpenAI's GPT-3 model, fine-tuned for use in programming applications. OpenAI has released an API for Codex in closed beta. Based on GPT-3, a neural network trained on text, Codex has additionally been trained on 159 gigabytes of Python code from 54 million GitHub repositories. You can find more information here https://openai.com/blog/openai-codex/

General
Relese date	August 31, 2021
Author	OpenAI
Repository	-
Type	Autoregressive, Transformer, Language model

Start building with Codex

We have collected the best Codex libraries and resources to help you get started to build with Codex today. To see what others are building with Codex, check out the community built Codex Use Cases and Applications.

OpenAI Codex tutorial: Natural language to SQL query with OpenAI Codex

👉 Discover more Codex Tutorials on lablab.build

Boilerplates

Kickstart your development with a Codex based boilerplate. Boilerplates is a great way to headstart when building your next project with Codex.

Codex Boilerplate Create a function just by typing what it should do, with help of OpenAI Codex.

Libraries

A curated list of libraries and technologies to help you build great projects with Codex.

OpenAI Codex Python Client Library The OpenAI Python library provides convenient access to the OpenAI API from applications written in the Python language.
OpenAI Codex NodeJS Client Library The OpenAI Node.js library provides convenient access to the OpenAI API from Node.js applications.

Supply Dispatcher

Supply Dispatcher is an autonomous AI agent designed to recover procurement orders when suppliers experience delays or fail to meet delivery commitments. The system analyzes supply disruptions, evaluates available alternatives, selects suitable replacement suppliers using deterministic decision logic, and orchestrates a structured recovery workflow. Its architecture combines intelligent decision-making, deterministic selection, isolated execution, validation, controlled approval, and verification of recovery outcomes. A key capability is the structured coordination of recovery operations, ensuring that decisions follow defined criteria and that execution results remain traceable and reproducible. The system supports procurement operations, supply chain management, supplier disruption recovery, and operational resilience. It aims to reduce manual intervention, improve recovery efficiency, and provide a systematic mechanism for responding to supplier failures. The project includes an automated testing suite with 333 passing tests, including 299 independent scenarios. Supply Dispatcher demonstrates the application of autonomous agent architecture, deterministic decision-making, and controlled execution to complex procurement workflows.

QuantGate

QuantGate helps ML engineers decide whether a smaller activation format is ready to release. A single prompt defines the dataset, encoded activation budget and maximum permitted accuracy loss. Vultr Serverless Inference plans the evaluation and generates a constrained Python scoring kernel. A disposable Docker worker on Vultr Cloud Compute compiles the kernel to native C++, verifies calibration results against an independent Python reference and returns measurable evidence. Before code generation, QuantGate checks exact sample overlap across data partitions. It freezes the candidate on development data before evaluating holdout slices. The synthetic sensor demo compares FP32, FP16, BF16 and INT8 with FP32 as the reference. INT8 uses 48 encoded activation bytes per sample versus FP32's 192, but its low-margin holdout accuracy falls to 59.38%, so the gate blocks release. The report shows source code, fingerprints, per-slice measurements and the execution trace. The sandbox disables networking, runs as a non-root user with a read-only root filesystem, and limits CPU, RAM and runtime. A timeout probe demonstrates worker termination and verified container removal. The agent explains the measured verdict after execution. Format rounding is simulated at layer boundaries, and byte counts cover encoded activation values rather than total process memory. The public demo also includes an Optical Digits classifier and an injected-overlap scenario.

SourceCourt Exact Lite

SourceCourt Exact Lite is a bounded, evidence-first research assistant built for AMD AI League Match 4. It answers source-grounded questions while refusing to invent certainty when the available evidence does not support an exact answer. The system retrieves and parses bounded HTTP/HTTPS sources, interprets a user’s question, identifies relevant evidence, ranks candidate passages and links, and extracts an exact answer only when the evidence is sufficient. When the answer is ambiguous, unsupported, or unverifiable, SourceCourt refuses rather than fabricating a result. The project was hardened against unsafe retrieval behavior. Its transport layer rejects private and loopback network targets, protects against SSRF and DNS rebinding, preserves TLS hostname verification, limits redirects, constrains response size and timing, and avoids unsafe proxy behavior. It also includes bounded resource use, deterministic error handling, citation checks, and explicit ambiguity/refusal behavior. SourceCourt Exact Lite is packaged as a real command-line product. It can be installed as a Python package, run locally, or executed from a container image. The public release includes reproducible setup instructions, a deterministic demo, and a security-focused test suite. The release is provider-neutral at the model-adapter boundary, keeping retrieval, ranking, evidence extraction, citation, refusal, and safety logic independent from any single model provider while allowing the official Match 4 runtime or grader to supply the required model configuration. Container: `ghcr.io/null007void/sourcecourt-exact-lite:1.1.2` Digest: `sha256:39cd2d31e6fb95219da190d12a5620868e68293ac6ad88f8659142b744e4feec` SourceCourt Exact Lite returns the exact answer when the source supports it, shows the supporting evidence, and refuses when the result cannot be verified.

AEGIS

AEGIS – Short Version AEGIS is a low-cost autonomous ground rover for border and field patrolling, target navigation, surveillance, and operation in GNSS-denied environments. Built on ROS 2 and Nav2, it supports GPS-based target navigation and multi-waypoint autonomous patrol missions. When GPS becomes unreliable, AEGIS switches to local navigation using odometry, inertial sensors, visual odometry, and pre-built maps, enabling continued patrol without satellite positioning. Its camera-based perception system uses YOLOv8 to detect drone-like aerial targets, people, and predefined human activities, generating operator alerts for suspicious events. It does not identify individuals or infer intent. The rover uses a rechargeable battery with solar-tracking support for longer field endurance. A modular, regulated RF counter-UAS research module is included for authorized defence/testing environments, focusing on GNSS-degradation detection rather than operational jamming. AEGIS is intended as a safe, modular technology demonstrator for perimeter security, disaster response, infrastructure monitoring, and search-and-rescue missions, reducing human exposure in hazardous areas..

GOA: Shop from a Sentence

Shopping starts with a need, not always a product name. Alice is graduating in Cape Town. She wants a green, elegant, reusable outfit for a warm December day, but conventional keyword searches leave her bouncing between shops. Independent designers remain hard to discover, while merchants cannot see what customers tried to find. GOA: Shop from a Sentence is a working creative-commerce prototype built by Busi Ndlovu and Kholofelo Robyn Rababalela for the AMD Developer Hackathon Act III, Track 3: Reinvent Commerce. A shopper describes an outfit in natural language or supplies reference photos. Gemma-4-31B-IT through AMD Token Factory interprets the brief into garment, footwear and accessory needs. Semantic catalogue matching proposes items; deterministic Python rules verify size availability, stock, arrival before the event and budget including delivery. Shoppers can swap pieces, ask why they fit, save a look and explicitly send demo quote enquiries. Unmet needs are shown honestly as "not on offer yet" and recorded in a maker demand-gap view. We evaluated GOA using 20 frozen briefs, 20 live Gemma calls and the same synthetic catalogue for a keyword-search baseline. GOA produced 55% fully valid looks versus 30%, 100% first-shown item grounding versus 82%, and 80% sendable enquiries versus 35%. Median steps were 3 versus 8, and GOA captured 12 unmet parts versus none for the baseline. Keyword search filled more individual parts (93% versus GOA's 83%), so our claim focuses on verified outcomes, not superficial coverage. The prototype runs on Vercel. Merchants, product inventory and quote inbox are fictional demo data; no purchases, actual merchant fulfilment or sales uplift are claimed. GOA demonstrates a more accountable bridge between human imagination and available supply.

OpenLoop

Industrial shift logs record what operators wrote, not what remains unfinished. Technicians leave free-text notes covering machine status, maintenance work, delayed reports, and corrections. When a note logs a pending safety test and a later note says the pump is operating normally, or a backdated entry arrives after a task was reopened, incoming supervisors must manually reconcile what is still open before the next operating window. OpenLoop is an evidence-based semantic reconciliation layer for industrial shift handovers. Instead of asking an LLM to summarize notes or maintain state in prompt memory, OpenLoop separates semantic extraction from state resolution: 1. Semantic Extraction (Gemma-4-31B-IT): Reads each unstructured shift note against the equipment and task registry and emits strict JSON containing typed assertions (pending, completed, status, retract), a mandatory verbatim substring quote, and a resolved occurrence timestamp via an OpenAI- compatible HTTPS adapter. Ambiguous equipment references are routed to supervisor review rather than guessed. 2. Deterministic Python Evidence Ledger: Verifies that every quote appears literally in the source note, isolates equipment operating status on a separate rail from task obligations, filters available knowledge by recording time (recorded_at), and resolves task state from the latest active assertion by occurrence time (occurred_at). Explicit corrections withdraw prior claims by ID without deleting historical records. Across two runs of a four-way benchmark on four synthetic held-out sequences (12 checkpoints per run), OpenLoop produced correct documented task states at 19 of 24 planned checkpoints (12/12 in Run 1; 7/12 in Run 2 with 5 client 45s timeouts) and zero false closures among its 19 valid checkpoints, while the full-history LLM baseline produced false closures at the same two checkpoints in both runs.