API reference

Reference data

Three endpoints with no parameters worth speaking of, backing a launch form: which origins exist, which models exist, and which pairs can be quoted against.

GET /origins#

Every Origin account on chain, ordered by id. No parameters. Origins are append-only and can never be edited or removed, so this list only ever grows.

GET /origins

{
  "source": "chain:Origin accounts",
  "items": [
    {
      "id": 1,
      "pubkey": "4LjMhsTWQrwqLPbHY6z5gf1CCxVennQcMSKmwTpfAw5g",
      "name": "stocks-only",
      "asset_class_mask": 79,
      "max_risk": 100
    }
  ]
}

The registry id. Also the PDA seed: ["origin", id].

pubkey

string

The origin account address, so a client can pass it to init_policy without deriving it.

asset_class_mask

u32

A bitwise OR of the asset class bits this origin permits. See risk, origins and asset classes for the bit values.

max_risk

The ceiling on Policy.risk for policies of this origin.

Field naming is snake_case here, and camelCase elsewhere

This endpoint returns the database row shape directly, so its keys are asset_class_mask and max_risk. The same concepts appear as assetClassMask and maxRisk inside /coins/:mint. That is an inconsistency in the API, not a difference in meaning.

GET /models#

Every Model account on chain, ordered by id. No parameters. Currently 21 entries across 9 providers.

GET /models

{
  "source": "chain:Model accounts",
  "items": [
    { "id": 1, "name": "GPT-5.5",       "provider": "OpenAI",    "enabled": true, "registeredAt": "2026-09-23T00:08:28.000Z" },
    { "id": 5, "name": "Claude Opus 5", "provider": "Anthropic", "enabled": true, "registeredAt": "2026-09-23T00:08:30.000Z" },
    { "id": 9, "name": "Llama",         "provider": "Meta",      "enabled": true, "registeredAt": "2026-09-23T00:08:32.000Z" }
  ]
}

u16

PDA seed: ["model", id], little-endian. This is the value written to Policy.model_id.

enabled

boolean

Whether the model can be picked at launch right now. A disabled model blocks init_policy and set_policy, and does not affect policies already pointing at it.

registeredAt

ISO 8601 string

Block time of the transaction that last wrote this entry, from the indexer. Not a field on the account. Because register_model overwrites in place, this is a last-changed time rather than a created time.

There is no price here, and there is not going to be. Prices live off chain and are documented on compute costs.

GET /quotes#

StonkFun's launchable pair catalogue, polled into the indexer.

Parameter	Type	Default	Effect
`search`	string	none	Case-insensitive substring match against `symbol` or `name`. Omit it to get the whole catalogue.

Ordered by policy count descending, then symbol ascending, and hard-capped at 600 rows. There is no offset.

GET /quotes?search=SPY

{
  "source": "stonkfun:/pairs + /launchlab/pricing",
  "items": [
    {
      "mint": "XsoCS1TfEyfFhfvj8EtZ528L3CaKBDBRqRapnBbDF2W",
      "symbol": "SPYX",
      "name": "SP500",
      "decimals": 8,
      "category": "xstock",
      "launchable": true,
      "launchLabReady": true,
      "logoUrl": "https://www.stonkfun.xyz/api/asset/quote-logo/XsoCS1TfEyfFhfvj8EtZ528L3CaKBDBRqRapnBbDF2W?v=c2b8ddb3",
      "policyCount": 0,
      "startMarketCapUsd": null,
      "graduationMarketCapUsd": null,
      "graduationMultiple": null
    }
  ]
}

Not every listed pair can actually launch

The endpoint reports launchable and launchLabReady as StonkFun reports them, and every one of the 509 rows read on 2026-09-23 is true for both. It does not report the mint's token program, and 222 of the 509 are classic SPL, which init_policy cannot pair. The catalogue moves, so re-derive that split rather than quoting it. See quote assets and pairs.

Where a quote's name comes from#

Elsewhere in the API a quote mint appears attached to a policy, and it needs a symbol. Two sources are tried in order, and the response always says which one it got.

Order	Source	Reported as
1	StonkFun's /pairs catalogue	`"stonkfun:/pairs catalogue"`
2	the quote mint's own Token-2022 metadata extension	`"chain:quote mint Token-2022 metadata"`
neither	nothing	`quoteSymbolSource: null`, plus a `quoteSymbolAbsentReason` naming both places that were checked

The absent case, verbatim#

GET /analytics, a quote with no first-party name

"quoteSymbol": null,
"quoteSymbolSource": null,
"quoteSymbolAbsentReason": "this mint is not in StonkFun's /pairs catalogue and its mint account carries no Token-2022 metadata, so there is no first-party name for it. Both sources were checked."

The launched token's own name, symbol and image come from the same Token-2022 metadata mechanism on the base mint, and are reported as mintName, mintSymbol and mintImageUrl on a ranking row. They are prefixed mint because a row already carries the quote leg's symbol and the two must not be confused.