ATOM MCP
The Model Context Protocol is an open standard for connecting AI assistants to external data sources. ATOM MCP makes our entire pricing dataset, plus the capability intelligence behind every model, available through this protocol, so your agent can search models, compare vendors, and rank options by capability and cost without leaving the chat environment.
Connect once with a JSON config block, then query in plain language. Your agent gains access to live pricing across the inference economy, with real-time pricing, capability metadata, and benchmark indexes. No API keys to manage, no rate limits to navigate, no separate dashboards to maintain.
Each tool addresses a different decision-making question. Free tier gives market-level intelligence and aggregated benchmarks. PRO tier unlocks SKU-level search, model deep-dives, vendor catalogs, and direct price comparison.
| Tool | What It Does | Tier |
|---|---|---|
| list_vendors | Returns all inference vendors tracked by ATOM, grouped by channel type (model developer, cloud marketplace, inference platform, neocloud). | Free |
| get_kpis | Returns the 9 ATOM Inference Market KPIs: output premium, caching discount, open-source discount, context window cost, model size spread, reasoning premium, platform discount, neocloud discount, and caching availability. | Free |
| get_model_intelligence | Returns the 6 Model Intelligence KPIs from model metadata: reasoning tier share, long-context saturation, frontier context ceiling, output ceiling spread, training cutoff lag, and vendor modality breadth. | Free |
| get_index_benchmarks | Returns the AIPI benchmarks (modality, channel, tier, special) with current values, weekly change, and YTD performance. | Free |
| get_market_stats | Returns aggregate market intelligence: total SKUs, models, vendors, modality coverage, and geographic distribution. | Free |
| search_models | Search models across all vendors with filters for context window, tool calling, modality, parameter count, and pricing range. | PRO |
| get_model_detail | Deep-dive on any model: full pricing across vendors, capability metadata, training cutoff, license, and historical price movement. | PRO |
| compare_prices | Side-by-side price comparison across multiple vendors for the same model. Returns input, cached, and output pricing with cheapest-first ranking. | PRO |
| get_vendor_catalog | Returns the complete SKU catalog for any tracked vendor with full pricing dimensions. | PRO |
Add the ATOM MCP server to your client's configuration. Authentication uses a Supabase-backed API key system with 5-minute caching.
{ "mcpServers": { "atom": { "command": "npx", "args": ["-y", "@a7om/mcp-server"], "env": { "ATOM_API_KEY": "your_key_here" } } } }
Once added, restart your client and the ATOM tools appear automatically. Free tier requires no API key. PRO tier uses the key generated when you subscribe.
ATOM MCP is also listed across the MCP ecosystem, including Smithery, the Official MCP Registry, MCP.so, npm, mcpservers.org, PulseMCP, Glama.ai, LobeHub, MCPMarket, and GitHub.
The free tier gives immediate access to ATOM's market intelligence. PRO unlocks the operational tools that let your agent make specific model and vendor decisions at runtime.
| Feature | Free Tier Free | PRO Tier $500/mo |
|---|---|---|
| Full vendor list | ✓ | ✓ |
| Market KPIs | ✓ | ✓ |
| Model Intelligence KPIs | ✓ | ✓ |
| AIPI Benchmarks | ✓ | ✓ |
| Aggregate market stats | ✓ | ✓ |
| SKU-level model search | — | ✓ |
| Model deep-dives | — | ✓ |
| Cross-vendor price comparison | — | ✓ |
| Full vendor catalogs | — | ✓ |
| Update frequency | Weekly | Weekly + Daily Price Alerts |
| Support | Email, 24h SLA | Email + Phone, 8h SLA |
