Developers
Mintok for developers & agents
Make live AI compute economics — model pricing in $/M-token, capability tiers, accelerator specs — callable by your agents and code. Current pricing is free; price history and investment BOMs are on the Analyst plan.
MCP server recommended for agents
Point any MCP client (Claude, Cursor, ChatGPT connectors) at the Mintok server and your agent can look up current model pricing and the cheapest capable model at call time.
Endpoint (Streamable HTTP): https://mintok.ai/api/mcp
{
"mcpServers": {
"mintok": {
"type": "http",
"url": "https://mintok.ai/api/mcp"
}
}
}Add an Analyst API key to unlock the keyed tools (price history):
{
"mcpServers": {
"mintok": {
"type": "http",
"url": "https://mintok.ai/api/mcp",
"headers": { "Authorization": "Bearer mk_your_key" }
}
}
}| Tool | Does | Access |
|---|---|---|
| list_models | Models with live $/M-token pricing, tier, size class, context (filterable). | free |
| get_model | One model by id (e.g. kimi-k3). | free |
| cheapest_model | Cheapest capable model — per tier, or ?tier=N. | free |
| list_silicon | Accelerators: specs + modeled cloud $/hr. | free |
| model_price_history | Historical $/M-token series for a model. | API key |
Free public feed
Unauthenticated JSON — the same current data the catalog pages render. CDN-cached.
GET https://mintok.ai/api/public/models # models + live $/M-token (filters: vendor, tier, sizeClass, maxOutputPrice)
GET https://mintok.ai/api/public/cheapest?tier=6 # cheapest capable model per tier
GET https://mintok.ai/api/public/silicon # accelerators + specs
GET https://mintok.ai/api/public/model-lineup # models + recent price movesAnalyst API API key
Key-authed REST for the moat data — historical price series, chip catalogue, and investment BOMs — with CSV export (?format=csv). Part of the Analyst plan.
curl -H "Authorization: Bearer mk_your_key" \
https://mintok.ai/api/v1/models/kimi-k3/price-history
GET https://mintok.ai/api/v1/models
GET https://mintok.ai/api/v1/models/{id}/price-history
GET https://mintok.ai/api/v1/chips
GET https://mintok.ai/api/v1/investment/architectures
GET https://mintok.ai/api/v1/investment/architectures/{slug}/bom
GET https://mintok.ai/api/v1/investment/price-bookFinOps ingestion — push your AI usage to attribute spend + get savings recs:
curl -X POST https://mintok.ai/api/v1/usage \
-H "Authorization: Bearer mk_your_key" -H "Content-Type: application/json" \
-d '{"events":[
{"model":"gpt-5.6-luna","provider":"openai","inputTokens":1200,"outputTokens":800,"team":"search","feature":"summarizer"},
{"model":"claude-sonnet-5","inputTokens":500,"outputTokens":300,"costUsd":0.012}
]}'
# costUsd is optional — omit it and Mintok prices the tokens from the catalog.
# Attribution tags (team, feature, environment) power the FinOps breakdowns.Get a key: Analyst plan or contact us.
Prices maintained by Mintok's catalog watchers from public sources; each record carries an asOf date. CORS-open, ~5-min CDN cache. Not investment advice. Methodology · Glossary.