Presskit
Boilerplate, quotable facts, naming conventions, and brand usage guidance for journalists and analysts writing about Sage AI.
Media contact
For interviews, product questions, fact-checking, or brand assets, email support@sageworksai.com with "Press" in the subject line. Tell us your outlet, your deadline, and what you need — we would much rather answer a question directly than have a detail guessed at. If you are on deadline, say so in the first line.
One-line description
Sage AI is a local-first AI coding assistant that runs open-weight models on your own hardware, so your source code never has to leave your machine.
Short description
Sage AI is a local-first AI coding assistant available as a command line tool, a web app, iOS and Android apps, and an SMS bridge. It supports more than 200 open-weight models — including the Llama, Gemma, Qwen, Mistral, Phi, DeepSeek-Coder, and Devstral families — running either as local GGUF weights on your own hardware or through anonymized cloud routing. Because prompts are processed locally by default, developers can use AI assistance on proprietary or regulated codebases without sending source code to a third party.
Longer description
Most AI coding assistants require the same trade: upload your codebase, get help. Sage AI is built to remove that trade. It installs with a single pip command, runs open-weight models directly on the developer's machine, and treats the cloud as an explicit opt-in rather than a default. When a developer does choose a cloud-hosted model, the request is anonymized before it leaves the machine — identifiers stripped, emails and phone numbers scrubbed — so the upstream provider receives a request that is not linked to a person.
The product is agentic: it reads and writes files, runs approved shell commands, and iterates against real test results rather than producing code in isolation. It integrates with the tools developers already use, and an SMS bridge lets a developer send a task to one of their own registered computers from a text message and receive the result back over the same channel. Sage is funded by subscriptions rather than data — user code is never used to train models, nothing is sold to third parties, and account deletion is immediate and permanent.
Facts you can quote
•
Product name: Sage AI. Website: sageworksai.com.
•
Install: pip install sage-ai-cli, then sage login.
•
Platforms: command line tool, web app, iOS app, Android app, and an SMS bridge.
•
Models: 200+ open-weight models, including Llama, Gemma, Qwen, Mistral, Phi, DeepSeek-Coder, and Devstral — as local GGUF weights or via cloud routing.
•
Integrations: VS Code, Cursor, JetBrains, Neovim, Sublime Text, Zed, GitHub CLI, Git Hooks, Godot 4, and the Unity Editor.
•
Plans: Free at $0/month (browser-only AI with 7 in-browser models), Starter at $19/month (300,000 output tokens), Pro at $49/month (2,000,000 output tokens), Premium at $99/month (10,000,000 output tokens).
•
Billing model: metered by output tokens, with a published per-1,000,000-token overage rate. Input tokens are not billed against the allowance.
•
Payment methods: PayPal, Apple Pay, and Google Pay.
•
Privacy position: prompts run locally by default; cloud requests are anonymized; user code is never used for training; deletion is immediate with no shadow copies.
How to write the name
•
Write it as "Sage AI" on first reference, and "Sage" thereafter. Not SAGE, not SageAI, not Sage.ai.
•
The command line tool is "sage" in lowercase, because that is the command you type. The package is sage-ai-cli.
•
"Local-first" is hyphenated.
•
It is an AI coding assistant, not an AI model. Sage runs models; it is not one.
Common mistakes worth avoiding
•
Sage does not train models. It runs existing open-weight models published by others, each under its own license.
•
Local-first does not mean cloud-only-if-you-pay. Local models are the default path on every plan that includes the CLI, and they keep working with no network connection at all.
•
The Free plan is browser-only with 7 in-browser models. It is not the same thing as running local models through the CLI.
•
Billing is on output tokens, not requests, not seats, and not input tokens.
•
Sage is not open source. The models it runs are open weight; the product itself is a commercial subscription.
Brand and logo usage
You may use the Sage AI name and logo editorially — in articles, reviews, comparisons, and coverage — without asking us first. Please do not alter the logo's proportions, recolour it, add effects to it, or combine it with other marks. Do not use our name or logo in a way that implies we endorse, sponsor, or partner with your product, and do not incorporate them into your own branding, domain names, or app icons.
For logo files, product screenshots, or anything else you need at print resolution, email support@sageworksai.com and tell us the format and context — we will send what fits.
Fact-checking
If you are writing about Sage and want a claim confirmed before you publish, send it to us and we will tell you whether it is accurate — including when the honest answer is that it is not. We would rather correct something before it runs than after.