Heidelberg AI Curriculum

Showing 601–640 of 902 terms in the glossary

Letter O22 terms

OLLAMA_HOST Flags: An environment variable that controls which network address Ollama listens on; set it to 0.0.0.0:11434 to allow other computers on your local network to connect.
OLLAMA_KEEP_ALIVE Flags: An environment variable that sets how long a model stays loaded in memory after its last use; the default is 5 minutes, but you can set values like 24h to keep it loaded longer.
on-request Concepts: An approval policy that lets Codex work freely within its sandbox but pauses to ask for your permission whenever it needs to go beyond those boundaries.
Onyx Cloud Concepts: The managed, hosted version of Onyx (cloud.onyx.app) — sign up and connect sources without running any infrastructure; billed per user.
Onyx Lite Concepts: A lightweight, chat-only deployment of Onyx that runs in under 1GB of memory — for testing or simple chat use, versus the full stack with vector indexes and background workers.
open source Core: Software whose code is published under a licence that lets anyone use, study, change and share it. You can run it yourself, but support and updates depend on its community or the company behind it.
Also known as: open-source, open-source software, OSS
Open WebUI Tools: An open‑source web interface you can install to talk with an AI model on your own server.
open weights vs open licence Access & cost: You can download an open-weight model, but its licence may still restrict commercial use — always check before shipping (some MiniMax and Llama terms do this).
open-source repos Components: Publicly available code collections that anyone can view and modify, which provide parts of the voice layer.
open-weight Access & cost Core: A model whose trained weights anyone can download and run on their own hardware. Its licence still decides what you may do with it, and some open-weight licences are stricter than open-source software licences.
Also known as: open-weight model, open weights, open model
1 other definition
open-weight · Access & cost
A model whose trained weights you can download and run yourself — free to run and private — as opposed to a closed cloud API.
openai Concepts Libraries Libraries & Code: A Python library originally made for OpenAI's cloud service that can also talk to any OpenAI-compatible server — including Jan's local server — using the same code.
3 other definitions
openai · Libraries & Code
The name of a Python library originally made for OpenAI's cloud API; LM Studio intentionally speaks the same format, so you can reuse this library to talk to your local models instead.
openai · Concepts
A Python library (also the name of the company) that provides a standard way to call AI chat APIs; Ollama supports the same interface so you can use this library with local models.
openai · Libraries
The name of the official OpenAI Python package; because OpenRouter uses the same API format as OpenAI, you can install this package and point it at OpenRouter instead, saving you from learning a separate library.
OpenAI compatibility Concepts: Cerebras's API speaks the same language as OpenAI's API, so code written for OpenAI can be redirected to Cerebras by changing just the base URL and API key.
OpenAI embeddings.create Commands: An API call that takes a list of text strings and returns a matching list of numeric vectors (3072 dimensions) representing their meanings.
OpenAI Realtime API APIs: An OpenAI service that streams audio to and from a model over a WebSocket, enabling live voice interactions.
OPENAI_API_KEY Files & config: The environment variable holding your model key. Local servers ignore it; cloud providers require a real one.
openclaw onboard Commands: The guided first-run setup wizard for OpenClaw that walks you step by step through connecting your AI model, configuring your gateway, linking chat channels (WhatsApp, Telegram, etc.), and optionally installing OpenClaw as a background service.
opencode Concepts: An open-source AI coding agent that runs in your terminal (or as a desktop app or IDE extension) and helps you read, write, and change code by chatting with a large language model.
OpenCode Zen Features: A paid AI model gateway maintained by the opencode team that offers a curated set of tested and verified models on a pay-as-you-go basis, so you can access quality-checked models without setting up separate provider accounts.
opencode-voice Tool: The name of the tool that provides local speech‑to‑text and text‑to‑speech functionality.
opencode.json Files & config: The main configuration file for opencode where you set your preferred model, provider API keys, tool permissions, and other project-level settings.
Operation Billing: One module action. Make meters usage by operations — now billed as "credits".
Outcome‑Focused Prompting Prompting: Writing a prompt that describes the desired final product instead of step‑by‑step instructions, letting the system decide how to achieve it.

Letter P18 terms

package Core: A bundle of ready-made code that others publish so you can install and reuse it in your own project. Install packages only from sources you trust.
Also known as: software package, packages
package manager Core: A tool that installs, updates and removes software packages for you, such as pip for Python, npm for JavaScript or Homebrew for macOS.
Pages Features: A publishing feature that turns any Perplexity research thread into a formatted, shareable article with a public URL you can distribute or embed.
pandas Concepts: A popular Python library for working with tabular data such as spreadsheets and CSV files; used in many data-analysis scripts.
parameters Size & speed Core: The numbers inside a model that were set during training. Their count, often written like 8B for eight billion, gives a rough idea of a model's size, memory needs and ability.
Also known as: parameter, model size
1 other definition
parameters · Size & speed
The adjustable weights a model learns during training; more parameters (e.g. 7B vs 120B, where B = billion) usually means smarter but slower and heavier to run.
pass/fail grid Tools: A table that marks each example as pass (green) or fail (red) after running the prompt.
PATH Files Logic: An environment variable that tells the operating system where to look for executable files.
Also known as: Path
1 other definition
Path · Logic
Conditional branching — different actions run depending on the incoming data.
pd.read_csv Concepts: A pandas function that reads a CSV (comma-separated values) file from disk and loads it into a DataFrame so you can analyse it with Python.
PDF File Types: Portable Document Format file that can be uploaded and searched inside a notebook.
PDFReader Commands: A tool from LlamaIndex that loads the raw text of a PDF file so it can be further processed into chunks.
per-token pricing Access & cost: Cloud models bill by the token, charged separately for input and output; a small model can be roughly 5× cheaper than a frontier one.
permission Concepts: A setting that controls whether opencode can use a particular tool automatically (allow), must ask you first (ask), or is blocked from using it entirely (deny).
Perplexity Computer Tool: A feature that breaks a user’s goal into sub‑tasks and runs the best AI model for each task in parallel.
Perplexity Max Concepts: A higher-tier subscription above Pro that adds features like Model Council, priority access to new models, and additional advanced capabilities.
Perplexity Pro Concepts: The paid subscription tier (around $20/month) that unlocks unlimited Pro Search, Deep Research, file uploads, model selection, and API credits.
Perplexity Profile Features: A saved set of custom instructions that automatically apply to every new conversation you start.
personal access token Concepts: A secret string you create on GitHub that lets a script act like you when accessing repositories.
Personal Avatar Features: A lifelike AI copy of you that can appear and speak in Synthesia videos.

OLetter O22 terms

PLetter P18 terms

Letter O22 terms

Letter P18 terms