# Arcmira

> Arcmira is an SF-based AI company and the search engine for the spoken web.

This repo and the `arcmira` npm package are the official TypeScript and JavaScript SDK plus the `arcmira` command line. The live agent index is on the website; fetch that first.

## Tasks

Use Arcmira to pull YouTube transcripts, find quotes with timestamps, research podcast sponsors, and monitor company mentions. Start with the README examples. Keep source links beside quotes, preserve the reported coverage window, and distinguish captions from Premium transcripts. For agent tasks, connect the MCP server; for an application, use this SDK.

## Canonical

- [Site llms.txt](https://arcmira.com/llms.txt): Fetch this first. Product, API, citation rules.
- [Docs index](https://arcmira.com/docs/llms.txt): Catalog of every docs page.
- [API reference](https://arcmira.com/docs/api-reference): Every operation with curl samples.
- [OpenAPI](https://api.arcmira.com/v1/openapi.json): HTTP JSON contract at https://api.arcmira.com/v1
- [SDK reference](https://github.com/arcmira/arcmira/blob/master/reference.md): Every SDK method with its request and response types.
- [MCP server](https://github.com/arcmira/mcp): Remote MCP server at https://mcp.arcmira.com/mcp. For agents, prefer it over the SDK.
- [Arcmira MCP](https://arcmira.com/mcp): What the MCP server does.
- [Python SDK](https://github.com/arcmira/python): `pip install arcmira`.
- [Developers](https://arcmira.com/developers)
- [Homepage](https://arcmira.com)
- [Pricing](https://arcmira.com/pricing)
- [Contact](mailto:hi@arcmira.com): hi@arcmira.com

## Use

- Install: `npm install arcmira`. Key: `ARCMIRA_API_KEY` or `new ArcmiraClient({ apiKey })`.
- Client: `client.entities.resolve` (best, suggested or ask; reads take the ids it returns), `client.transcripts.search` (GET /v1/search), `client.transcripts.get` (a Premium read uses credits from the plan, then the on-demand budget, and answers pending on 202 until ready), `client.mentions.list` and `client.recommendations.list` (entity_id required, auto-paging), `client.entities.momentum`, `client.channels.sponsors.list`, `client.monitors.create`, `client.monitors.entities.add` (follow ids, or exact names with a type, in a monitor) and the rest, all in reference.md. Dated reads take `after` (inclusive) and `before` (exclusive).
- CLI: `npx arcmira setup` connects the MCP server and the Arcmira skills to Claude Code, Codex, Cursor, VS Code and Gemini CLI. `npx arcmira search "agent payments"`, `resolve` (`--context` for the user's words), `mentions`, `momentum`, `sponsors`, `recommendations`, `episodes`, `transcripts get|quote` (`transcript` is short for `transcripts get`), `occurrences`, `status`, `trackers create|list` (`follow` is short for `trackers create`, which takes `--monitor`), `monitors list|create|update|trackers|add|attach`, `integrations slack`, `login`, `logout`, `whoami`, `auth token`; any endpoint with `arcmira api GET /v1/...` (gh api grammar, `--paginate`), its parameters with `arcmira schema <command|operationId>`, docs search with `arcmira docs <query>`, worked tasks with `arcmira examples`. Filters take ids (`ent_`, `UC`); a name exits 2 with the resolve command. `--json` for the raw response, errors as JSON on stderr; exit 2 on bad input, 1 on API errors, 4 while Premium transcribes.
- Errors: typed classes under the `Arcmira` namespace with `err.body.error` (`type`, `code`, `gate`, `unlock.url`, `details.quote`).

License: Apache-2.0.
