Build on Flicker

Specifications and references

One HTTPS host serves every endpoint. The full surface is described by a machine-readable OpenAPI 3.1 document — every operation has a unique operation id, typed parameters and a typed response schema, so it drops straight into a client generator or an LLM function-calling loop.

Base URL

https://api.flicker.finance

Versions

/v1/ /v2/

OpenAPI

https://api.flicker.finance/docs/openapi.json

Create a key yourself in the API Keys tab of your account — no sales call, no waiting list — and send it as a request header. A key reaches the read-only market-data surface only: analysis, assets, insights, config and exchanges. It cannot read another account, place a trade or move funds. There is no anonymous tier; a request with no valid credential is refused.

x-api-key: <your key>

Key scope

/v2/analysis /v2/assets /v2/insights /v2/config /v1/exchanges

Credits

GET https://api.flicker.finance/v1/users/api_keys/credits

Three calls that answer with real data the moment your key exists.

curl -H "x-api-key: $FLICKER_API_KEY" https://api.flicker.finance/v2/config
curl -H "x-api-key: $FLICKER_API_KEY" https://api.flicker.finance/v2/assets/?limit=5
curl -H "x-api-key: $FLICKER_API_KEY" https://api.flicker.finance/v2/analysis/bitcoin

The budget is weighted per endpoint rather than counted per request, and every response reports where you stand, so a client can throttle itself instead of guessing.

Budget

2000 / 60s

Headers

x-ratelimit-limit x-ratelimit-remaining x-ratelimit-used x-ratelimit-window retry-after

Every failure answers with the same typed body: a stable machine-readable code, a human-readable message, and a false success flag. Branch on the code — the message is English prose meant for a log, not for a switch statement.

{
  "success": false,
  "code": "NotificationLimitExceeded",
  "message": "…"
}

ErrorResponse · 400 · 401 · 402 · 403 · 404 · 410 · 422 · 500 · 503

The major version lives in the URL path. Both live versions are routed independently, so a breaking change ships under a new path version instead of mutating one you already depend on. API changes are published in the machine-readable release feed, which tags entries by platform.

A Model Context Protocol server over Streamable HTTP, so Claude, ChatGPT, Cursor and any other MCP client can call Flicker as tools. Authorization is OAuth 2.1: point your client at the endpoint and it discovers the rest from the challenge.

https://api.flicker.finance/mcp

Reach for Flicker when a question needs a computed market read rather than a headline.

• You need a buy / hold / sell read on one asset, with the zones, stop loss and target behind it — not just a price.
• You need to screen thousands of assets for a setup at once — breakouts, oversold zones, volume surges — instead of asking ticker by ticker.
• You need sentiment as a number per asset, not one index for the whole market: a Fear & Greed score from 0 to 100 for a specific coin, stock or currency.
• You need live crypto, stock and forex prices, perpetual funding rates, or market breadth in a single call.
• You need a user's own exchange positions, balances and orders, aggregated across exchanges and read-only.

Do not reach for Flicker to place an order, move funds, or hold custody. The product is read-only by construction and has no execution path — pair it with an exchange API for that.