ATOM MCP server

Which MCP servers are ATOM's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.

read 2026-10-23

← Back to ATOM

Whose server it is

Registry nameio.github.AtticStandard/atom-mcp-serverpublished by AtticStandard on GitHub

Published under the GitHub organisation AtticStandard, the same one atticstandard.com links from its homepage. The registry ties that namespace to a GitHub sign-in.

When we asked itchecked 2026-10-23

Stateanswerschecked 2026-10-23
Tools11
Server calls itselfattic-standard-mcp
Server version2.1.0
Protocol version2025-06-18

On 2026-10-23 the server answered initialize and tools/list without a sign-in. First answer in 82 ms.

Should an agent connect itNot proven either way yet

Whether its tools do the product's job has not been judged yet.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

Measured dailydaily checks: 11, the first on 2026-10-23

INITE MCP score86 of 100answers 100 · definitions 100 · context 96 · use 0 · connects 100
Answered100% of days11 of 11 daily checks
Median first answer81 ms
Space in an agent's context≈ 3,440 tokensthe definitions of all its tools (11), which an agent carries on every turn
Tool definitionsA · 100 of 100
Typed results0 of 11 toolsdeclared outputSchema: results come typed, read once

Every tool passes all six checks.

What an agent can do with it

not sure · settle inference invoices against a weekly benchmark level
can do · compare a company's model list price with the market median
not sure · review which vendor repricings moved a benchmark each week

Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?

Worth a look

None of the five warning signs in these definitions.

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

Tools it offers11 tools · read 2026-10-23

get_index_benchmarks
The Attic Standard price indexes for AI inference: every published index at the week the site shows. Six families: Modality (text, multimodal, image, video, audio, voice, embeddings), Channel (model developers, cloud marketplaces…
get_price_history
Weekly history. With index_code: the index series (benchmark level, May 2026 = 100, and the spot price with its 25th and 75th percentiles) for up to 260 weeks; free. With model_name: every vendor's price for that model week by week…
get_kpis
The nine market KPIs Attic Standard publishes each week. - Price structure: output premium, caching discount, caching availability - Price dynamics: repricing activity, repricing depth, post-launch drift - Competition: multi-vendor…
get_model_intelligence
Six capability and coverage measures drawn from the metadata behind every tracked model: reasoning tier share, long-context saturation, frontier context ceiling, output ceiling spread, training cutoff lag and vendor modality breadth. Read…
get_index_constituents
What sits inside an index basket this week. Free tier: the composition (SKUs, models and vendors, split by channel, origin, tier and license). Attic Standard MCP PRO: the model-by-model basket with the vendors selling each model…
get_market_stats
Coverage of the tracked market (active vendors by channel, countries, models, priced SKUs, published indexes) and price distributions split by modality, unit and direction. Free tier returns counts, ranges and redacted samples; Attic…
list_vendors
Every vendor in the Attic Standard fleet with its channel (model developer, cloud marketplace, inference platform or neocloud), country, region and pricing page. Filter by channel, region or country.
search_models
Search every priced SKU by modality, vendor, channel, creator, family, tier, license, origin, reasoning, open weights, direction, maximum price and minimum context window. Cheapest first. Free tier returns counts, ranges and redacted…
get_model_detail
One model in depth: specs (creator, origin, tier, license, context window, output limit, training cutoff, modalities), the Attic Standard indexes it belongs to, and its price at every vendor. Free tier returns counts, ranges and redacted…
compare_prices
The same model, or a whole family, priced across every vendor that sells it: cheapest and dearest offer, vendor count and spread ratio for each direction, plus each offer with its channel. Free tier returns counts, ranges and redacted…
get_vendor_catalog
Everything one vendor sells: channel, country, pricing page, model and SKU counts, and the full price list. Free tier returns counts, ranges and redacted samples; Attic Standard MCP PRO ($500/month, https://atticstandard.com/mcp) returns…

How to connect

Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:

{
  "mcpServers": {
    "atom-mcp-server": {
      "type": "http",
      "url": "https://mcp.atticstandard.com/mcp"
    }
  }
}

Add it

Claude Code, in a terminal

claude mcp add --transport http atticstandard https://mcp.atticstandard.com/mcp

Add to Cursor Add to VS Code

claude.ai, ChatGPT and other clients: add https://mcp.atticstandard.com/mcp as a custom connector in the settings.

Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.