"OCR agent" matching MCP tools:

• fetch_url

  Heista

  Drill into a specific URL after search surfaces it. Returns the extracted text content plus metadata. Internal routing: PDFs hit Anthropic Files API for OCR + structured extraction; HTML pages are fetched + text-extracted via readability-style stripping. Use for: verifying a verbatim quote from a Reddit thread, reading a primary source in full (earnings transcript, research paper), drilling into a vendor product page after search surfaced the URL. NOT for: discovering new URLs — use search/search_community/search_research first. This tool takes a known URL only. Optional max_chars 100-50000, default 8000. SSRF-protected: private IPs + localhost blocked.

  Connector

• extract_contract_from_url

  transaction-coordinator

  Extract structured transaction data from a contract at a URL. Downloads the document, extracts text (with OCR fallback for scanned PDFs), and runs PrimaCoda's contract-extraction prompt to return parties, addresses, dates, prices, and key contract fields. Use this when an agent has the contract hosted somewhere (Dropbox, Google Drive direct download, Square Space, etc.) and wants to skip the upload step. For multi-document deals (purchase + addenda + disclosures), use the PrimaCoda dashboard's batch upload — this tool handles ONE document. Args: pdf_url: Direct download URL for the contract (PDF, DOCX, TXT, or image). Must be reachable from the PrimaCoda server. Google Drive "shared link" URLs work if set to "anyone with link"; other share URLs may need their direct-download form. api_key: Your PrimaCoda MCP API key (starts 'pck_').

  Connector

• get_article

  Islam West Africa Collection (IWAC)

  Get one article (by id): full metadata, the AI abstract (description_ai), AI sentiment, and OCR text. Pass a `keyword` to get ~2000-char excerpts around each match instead of the full (capped) OCR.

  Connector

• agent-brain

  MeshMarket

  Agent Brain — Reason over a question or task with your agent's own persistent memory in the loop: recalls up to 12 relevant memories from your agent's private scope, reasons with Claude, and writes up to 3 new memories back, so the agent improves with every call. Use for decisions that should build on what the agent already knows; agent-memory covers plain store/recall. Input: {think: string}. Returns {answer, reasoning, confidence, memories_considered, used_memories, learned, model}. (8 MESH/call, a tool · cognition)

  Connector

• get_opinion

  Court Listener

  Get the FULL TEXT of a US court opinion by its CourtListener opinion ID (the opinion_id returned by search_opinions). Covers 8.3 million opinions, 71% state case law — no API key needed. Returns the opinion text, the case name, the real filing date, author, citation count, whether the text came from OCR of a scan, and the data as-of date. Opinions filed after that date may fall through to live search.

  Connector

• document.extract_text

  api

  Extract plain text from a PDF or image (base64-encoded). Use when you need raw text for downstream AI analysis (summarization, claim checking, structured extraction). For documents at a public URL, use url.extract instead (no base64 encoding needed). Returns: { pages: number, text: string } Example prompts: - "Extract the text from this scanned contract so I can search it." - "Give me the raw text from this PDF document." - "OCR this image and return the text content."

  Connector

• get_extract

  webbersites-x402-data-api

  Document extraction: fetch a PDF, DOCX, or CSV by URL and get clean Markdown plus structured JSON — PDF text by page with metadata (honestly flags scanned PDFs that would need OCR), DOCX converted to real Markdown, CSV parsed to typed columns + JSON rows + a Markdown table. For agents that need document contents, not bytes. ($0.02 per call, paid via x402)

  Connector

• recognize_font

  Jinero

  Identify which font is used in an image. Powered by our OWN CNN embedding model, trained on the jinero font catalog — it matches fonts by visual shape/style, so it needs NO OCR and NO text (works for Latin and Cyrillic). Send a tight crop of one line of text as either image_url (public URL) or image_base64 (base64/data-URI, e.g. a local screenshot). The image is processed in memory and deleted immediately — never stored. Returns the most visually similar font families with scores. Fast (~200 ms).

  Connector

• formation_extract_rsa_terms_ocr

  Lovie Company Formation

  Runs vision OCR on an already-uploaded signed Restricted Stock Purchase/Award Agreement (RSA) PDF for the active company and returns the founder and company names plus structured terms (total_shares, unvested_shares, price_per_share, grant_date, vesting_start_date / vesting_total_months / vesting_cliff_months, acceleration_clause, repurchase_right, 83(b) status, etc.) WITHOUT persisting anything. Empty string / 0 / false means the value was not stated in the document — never fabricate. Flow: first call GetOcrUploadURL with kind=RSA and upload the PDF, then call this tool with the returned source_s3_uri. To persist the grant and link the document, pass the extracted terms and the SAME source_s3_uri to CreateCapTableAgreement with type=CAP_TABLE_AGREEMENT_TYPE_RSA — preserve the vesting fields, do not drop them.

  Connector

• doc_parse

  stagenth · 文档解析

  把 PDF / Word(.docx) / PPT(.pptx) / HTML / 纯文本 解析成干净 Markdown + 结构化（按实际解析页数计费，每页 1 credit）。 纯解析 + 格式转换、不生成任何 AI 内容；扫描件 OCR 暂未支持。 支持 page_range 只解析指定页码段（只按解析的页数扣费）。

  Connector

• pair-agent

  Crinkl Commerce MCP

  Start the human-authorized agent pairing flow. Generates a 4-character code that a human approves in the Crinkl PWA to authorize the agent. Returns the code and expiration time (10 minutes). No authentication required.

  Connector

• read_pdf

  docweave-mcp

  Read a PDF and return its text as markdown (or plain text). Accepts a public URL or base64 bytes. Extracts the embedded text layer; a scanned, image-only PDF returns a needs-OCR notice instead of empty text. Priced per document; retries with the same idempotencyKey never double-read. The canonical way for an AI agent to ingest a document's contents.

  Connector

• usage_stats

  agent4.io

  Usage statistics. group ∈ total | agent | user | space | day | model. Metadata only — token counts and event counts, never any conversation content. Use group="user" to see who uses the most, group="agent" for the busiest agent.

  Connector

• convert_document

  Mdkit

  Convert a document to Markdown synchronously (the fast lane). Decode ``content_base64`` (the raw file bytes, base64-encoded) and run markitdown over it, returning ``{markdown, meta}`` where ``markdown`` is the converted text and ``meta`` carries the source ``filename`` and the output ``length`` in characters. Best for small office/HTML/text files; for large or complex documents (or OCR-heavy PDFs) use ``submit_conversion_job`` instead.

  Connector

• search_pages

  reading-room

  Literal search over the OCR text of all 5,247 pages, before segmentation into records. This is the completeness layer: it reaches duplicate copies and pages no record covers, and matches across a whole page rather than within a chunk. Use it to settle whether a phrase appears in the corpus at all — a miss here is much stronger evidence of absence than a miss in search_records, though OCR error and page-spanning lines still make it imperfect.

  Connector

• invoice.create

  DPX — Institutional Cross-Border Settlement

  Create an agent-to-agent invoice. Agent A calls this to request payment from Agent B. Returns an invoiceId and payUrl — Agent B calls invoice.pay with the invoiceId to settle. Invoice expires after ttlSeconds (default 24h).

  Connector

• get_sample

  extract

  Free recorded sample response for a paid product (from real production runs, marked sample:true) - see the exact response shape before paying anything. Available: ocr, enrich, verify, structured, jp-company, document-pipeline, pdf-info, meal-vision, portrait-analysis.

  Connector

• list_inbound_forwarding_addresses

  mailbox

  List the renter’s private inbound forwarding aliases on forward.mailbox.bot. These are the unique intake email addresses an operator, assistant, provider, or external agent can forward scans, PDFs, photos, provider notices, notes, and other context-aware documents to so mailbox.bot can build OCR-backed inbound context. Forwarding/emailing attachments here initiates OCR/extraction; this tool discovers the address and does not upload files directly into OCR. The alias is member-scoped, so live and sandbox agent keys for the same member resolve to the same intake address.

  Connector

• x711_substrate_leaderboard

  x711 — Universal Agent Gas Station

  Agent-seeded lifeforms ranked by fitness score. See which AI agent ideas survived the SUBSTRATE evolution engine. FREE — no credits required (counts toward 10/day free tier). No API key needed.

  Connector

• invoice.companyProfile

  Torify — Japan Locale APIs for AI Agents

  One-call composite: T-number (T + 13 digits) → invoice registration status (NTA 適格請求書 API) + corporate details (NTA 法人番号 API). 日本語: T番号 1 つでインボイス登録状態と法人詳細を同時取得。**Use when an AI agent receives a T-number (e.g. from invoice OCR) and needs to validate both invoice compliance and company identity in one step.** Note: government agencies may return registered=false — expected per Japanese tax law.

  Connector