The public radar exposes a current, read-only view without requiring an account or API key.

AI Resource Radar is a local-first tracker for free AI tiers, GPU compute, grants, and prices. It keeps the source and verification time beside each result, so the answer is practical: what is free, how much you get, when it resets, what the restrictions are, and how to claim it.

The default collection pipeline is deterministic: no AI, API key, cookie, or account data is required. Optional reviewable features remain isolated from collection and cannot change its verified evidence.

Start with uvx

Try the keyless collector without creating a checkout or virtual environment:

uvx ai-resource-radar start --open

For the hosted snapshot, use Live Radar or fetch the documented public data manifest. The public site is an aggregate view; always follow an official source before relying on an offer. Its v0.8.0 snapshot adds six bilingual scenario pages, stable Atom/RSS feeds, crawlable Chinese and English provider pages, conservative copyable integration examples, and on-demand catalogue loading while remaining static and read-only. Pages binds each build to a fresh 23-source refresh and Git commit.

Public scenarios, feeds, and search visibility

The hosted site publishes six read-only scenario pages in both languages at /zh/scenarios/<slug>/ and /en/scenarios/<slug>/. They turn common intents into short paths through the public catalogue; they do not create an account, personalize a result, or promise that an offer is still available.

The public change/resource stream is available as Atom at /feed.xml and RSS at /rss.xml, with English equivalents at /en/feed.xml and /en/rss.xml. Entries contain stable hash IDs, bounded public facts, observed times, and safe same-origin or official-source links. They are a convenient read-only snapshot, not a guarantee of quota, price, eligibility, or availability. Use GitHub Watch for release and maintenance updates; it is not a substitute for the daily resource feeds.

Production Pages can opt into Google Search Console site verification. The builder reads the repository secret only through AI_RADAR_GOOGLE_SITE_VERIFICATION_TOKEN and renders it as Google's required public verification marker in the homepage HTML. It omits the token from the manifest, JSON/XML feeds, logs, and workflow output. Local builds keep --search-console-provider none by default.

The 30-day discovery experiment starts at experiment_started_at in the public manifest. Its success gate is intentionally modest: at least 10 of 12 scenario pages indexed, exposure in three intent categories, and at least one click or bounded confirmation signal. This is an indexing and reach check, not an exact conversion measurement: a confirmation does not prove a signup, purchase, or other user-level conversion, and no user identity is exported.

What you get

Free tiers and AI prices change frequently, while ordinary link lists quickly become stale. This project turns public source material into a small, explainable local database:

Provider profiles keep the policy, evidence, verification time, and compatible integration examples together.

Quick start

Requires Python 3.11 or newer. Install from PyPI in an isolated environment:

python3 -m venv .venv
source .venv/bin/activate
python -m pip install ai-resource-radar

ai-radar refresh
ai-radar dashboard --open

The dashboard is available only on 127.0.0.1:18766.

The local Dashboard adds deeper filtering, claim steps, change history, and diagnostics without exposing local data.

On macOS, install the dashboard, menu bar helper, and 08:00 daily job:

ai-radar service install
ai-radar service status

To uninstall the services without deleting the database:

ai-radar service uninstall

Platform support

Windows is not tested yet. Linux CI verifies the deterministic core and dashboard; macOS CI also compiles and tests the Vision OCR and menu bar helpers.

What it tracks

The built-in adapters currently cover 23 sources:

Community sources can discover candidates but cannot upgrade an offer to “officially verified”. Each HTTPS source is allow-listed, limited to 16 MB, isolated on failure, and supports ETag/Last-Modified caching.

Explainable ranking

The radar deliberately avoids an opaque score:

Within a tier, results are ordered by mainland availability, estimated value, and recent changes. The dashboard shows the reasons instead of hiding them inside a number.

Optional extensions

AI efficiency tips

Official guidance and manually imported articles remain candidates until a human approves them. Approval is limited to the marked managed block in an AGENTS.md, creates a private backup, and is auditable and reversible. See docs/TIPS.md for the safety model and ai-radar tips --help for the CLI.

How it works

flowchart LR
    A[Allow-listed public sources] --> B[Source-specific deterministic parsers]
    B --> C[Normalized SQLite schema]
    C --> D[Explainable ranking and change detection]
    D --> E[Dashboard, CLI, and local notifications]

One failed source never clears data from other sources. A missing offer is removed only after two successful parses both confirm its absence. Parser drift keeps the last trusted value and marks the source for verification.

Useful commands

# Refresh all due sources, or bypass cadence
ai-radar refresh
ai-radar refresh --force

# Browse verified, no-card resources
ai-radar list --verified-only --no-card
ai-radar list --kind gpu --no-card

# Review recent changes
ai-radar changes --days 30

# Run the complete daily workflow
ai-radar daily

# Diagnose the database, source freshness, helpers, and services
ai-radar doctor
ai-radar doctor --json

Run ai-radar <command> --help for every filter and option.

Data, privacy, and storage

• SQLite schema v7 uses file mode 0600; no secrets, cookies, or account data are stored.
• Public scenario pages and Atom/RSS feeds contain only bounded, aggregate source facts and stable hash IDs; they never include local database IDs, request headers, cookies, or account data.
• Google Search Console verification is a production build concern only. Its token is read from the environment and rendered only in the required homepage verification marker; it is omitted from manifests, data feeds, logs, and workflow output. The Pages workflow never prints it.
• Tips retain only bounded summaries and evidence. Approval updates only the marked AGENTS.md managed block and creates a private backup under ~/.codex/backups/ai-tips/.
• Full fetched pages are parsed in memory and are not archived.
• Fetch logs are retained for 90 days; ordinary changes and delivered notifications for 365 days.
• Important free-tier changes and unread notifications are retained.
• Periodic cleanup and threshold-based VACUUM prevent unbounded growth.
• The dashboard accepts only loopback Host/Origin requests and serves no remote assets.

See Architecture and Security for details. Existing users should also read the v0.2 migration guide; complete removal is documented in Uninstall.

Development

git clone https://github.com/ai-resource-radar/ai-resource-radar.git
cd ai-resource-radar
python3 -m venv .venv
source .venv/bin/activate
python -m pip install -e .

python -m unittest discover -s tests -p 'test_*.py'
node --check src/ai_resource_radar/web/ai-resources.js

Contributions are welcome. Start with CONTRIBUTING.md, or open an issue with the official source URL and the policy or price that needs attention.

License

MIT