# Audioscrape > Audioscrape is the audio intelligence platform for AI agents: a searchable, speaker-attributed record of spoken audio, covering public podcasts, interviews, talks, and public meetings (a catalog of 18M+ episodes, over 140,000 hours transcribed) plus the customer's own uploaded calls, meetings, and recordings, served to agents over MCP and REST with named speakers, timestamps, and citations. ## What it is - Public corpus: a catalog of 18M+ episodes findable by title and show; over 140,000 hours (more than 200,000 recordings) transcribed, speaker-attributed, and searchable, growing daily. No crawling, downloading, or processing needed. - Private audio: upload calls, meetings, interviews, and recordings to a private workspace; transcribed with speaker separation and searchable alongside the public corpus. - Persistent speaker identity: name a speaker once; later recordings in the workspace are matched to that voice and named when the match is confident. Stateless transcription APIs start from zero on every file. - Every result carries speaker, timestamp, and a link to the source audio. - Also: keyword and semantic search, entity extraction and knowledge graph, alerts on spoken text (email/webhook), podcast charts for 31 countries, shareable excerpt links, multi-tenant workspaces with audit log and data export. ## What it is not - Not a per-request speech-to-text API and not real-time streaming; transcription is batch, the product is the searchable record. - Not a telephony or call-center integration. - Not a general web search engine; the index covers spoken audio only. ## Access - MCP server: https://mcp.audioscrape.com (OAuth 2.1 + PKCE, dynamic client registration). Works with Claude, ChatGPT, Cursor, and any MCP client. Setup: https://www.audioscrape.com/docs/integrations/mcp - MCP tools: search, search_audio, search_speakers, search_entities, search_podcasts, get_transcript, get_episode_overview, get_speaker, get_entity_graph, get_trending, get_charts, list_podcast_episodes, list_recent_transcripts, list_my_datasets, list_my_uploads, transcribe_audio (direct link, Drive/Dropbox/OneDrive share link, or chat attachment via file input), request_upload_url (one-time PUT URL for chat attachments), get_transcription_status, create_share_link, fetch - REST API: https://www.audioscrape.com/api/ (Bearer API key, date-pinned versioning). OpenAPI: https://www.audioscrape.com/api/openapi.json Swagger: https://www.audioscrape.com/api/docs/swagger Guide: https://www.audioscrape.com/docs/api - Without an account, a visitor can run one search on the website and browse public episode pages. API and MCP need a free account (Google, Microsoft, or a one-time email sign-in link; no passwords), or a pay-as-you-go workspace the agent opens itself. ## Pricing (mirrors https://www.audioscrape.com/pricing) - Free: $0, 300 transcription minutes/month (recordings up to 3 hours each), 100 searches/month (web, REST, MCP), 2 GB storage, no card, does not expire. REST data endpoints (libraries, items, transcript export) need a paid plan. - Starter: $35/month or $350/year, 1,200 transcription minutes (20 hours, any file length), 500 searches, 5,000 API calls per month, 25 GB storage. - Pro: $129/month or $1,290/year, 6,000 transcription minutes (100 hours, any file length), 5,000 searches + 2,000 semantic searches, unlimited API calls per month, 100 GB storage. - Enterprise: custom contract, priced per workspace (multiple dedicated workspaces under one contract, e.g. per team, client, or environment), 25+ members per workspace, contractual availability SLA and support terms, US or EU/EEA region pinning, DPA, security pack, custom volumes and overage schedule. https://www.audioscrape.com/enterprise - No free trials of paid plans; checkout charges immediately and can be cancelled any time. ## Pay as you go (autonomous agents, no sign-up) - Price list (machine-readable): GET https://www.audioscrape.com/api/agent/v1 - Paid answers reply HTTP 402 until paid: POST /api/agent/v1/search, GET /api/agent/v1/excerpt?episode_id=&t=, GET /api/agent/v1/overview?episode_id= - Top up a pay-as-you-go balance ($5 to $100, Machine Payments Protocol with a Stripe card or Link) at POST /api/agent/v1/credits; each paid call is debited from it. A first top-up without a key opens a pay-as-you-go workspace and returns its API key (send Authorization: Bearer sk_...) and a claim link for the person you work for. - Subscribers' agents use their workspace key and are served within the plan. - Transcription: POST /api/agent/v1/transcribe with a JSON body holding the file's url and a workspace key; poll GET /api/agent/v1/transcriptions/{job_id}. Pay-as-you-go pays per audio minute from the balance, charged at the measured length before transcription starts and refunded if it fails. - Paying accepts the Terms of Service (https://www.audioscrape.com/legal/terms) and Privacy Policy (https://www.audioscrape.com/legal/privacy). - No result, no charge. Answers, not records: at most 10 segments per search, 3 minutes or 6,000 characters per excerpt. - People and teams keep the plans above: https://www.audioscrape.com/pricing ## Trust - Uploads are isolated per workspace by the database itself on every query; only workspace members and agents holding their keys can search them. - Customer audio and transcripts are never used to train models. - Zero standing operator access: staff cannot read a workspace without a support request the customer approves; access is read-only, time-boxed, and written to the workspace audit log. - US or EU/EEA region pinning per workspace, on Enterprise contracts. SOC 2 Type I: audited as of 23 September 2026, report pending. DPA for enterprise. Owners export workspace data and close the account themselves. ## Key pages - Facts page for agents: https://www.audioscrape.com/docs/for-agents - Pricing: https://www.audioscrape.com/pricing - Docs: https://www.audioscrape.com/docs - MCP setup: https://www.audioscrape.com/docs/integrations/mcp - Search: https://www.audioscrape.com/search?q= - People, organizations, topics, places (the knowledge graph): https://www.audioscrape.com/graph - Trending podcasts (charts by country): https://www.audioscrape.com/podcasts ## Contact Audioscrape Inc. lukas@audioscrape.com https://www.audioscrape.com