MemeStack MCP Server

Search charts, graphs, maps and infographics with their source data: the dataset and its citation, the unit, the regions plotted and the period the chart actually shows. Two kinds of stored metadata: charts from Our World in Data and three statistical publishers (provenance 'authoritative') carry the publishing dataset and its citation, the period the chart displays (never the dataset's wider history), and whether the source has published newer data since the image was captured; community-uploaded charts (provenance 'extracted') carry a title, source line, unit and period read from the image itself and not checked against the original data. This tool covers only images that HAVE stored metadata, so use search_images for the widest chart coverage; search_images, get_image and cite_image carry the same metadata whenever a row exists. Set publisher_only to keep only the publishers' own charts (licence-clean, citable). Every image carries rights (source publisher/creator/community, licence, attribution line, commercial_use yes/no/unknown - unknown is never a guess), and match_quality says whether the search found any strong match (strong / closest_only / none). When corrected_query is present, the results are for that corrected wording, not the words as typed - say so to the user. Results exclude images classified controversial unless include_controversial is true.

Get ready-to-paste citations for one or more MemeStack images (memes, charts, screenshots): markdown, HTML and plain text naming the source and licence, plus a combined block for citing several images at once. Up to 50 ids per call.

Generate a new image from a text prompt using AI (Grok Imagine). Paid tool - billed per call via the agent payment rails (x402/L402); not covered by any free quota. Generation runs asynchronously on the server and this call polls until it completes, up to ~90 seconds. If you get a "still running" error, call the tool again with the SAME payment proof to resume polling - you will never be charged twice for one payment. Price: $0.08 standard or $0.20 quality, paid in sats over Lightning at the current BTC/USD rate or in USDC on Base. Note for x402 clients: "quality" mode exceeds the agents-SDK default per-call cap of $0.10 - raise maxPaymentValue in withX402Client to use it.

Get one image's full record by ID - a meme, chart, infographic or screenshot: caption, alt text, tags, the words printed in it (OCR text), its source and licence (rights), a ready citation and URLs. For a chart that is part of a series, chart.series lists every edition (other years) by id. When the image's receipts were already read, they ride along as receipts (see get_receipts).

Search memes, reaction images, charts, graphs, infographics and screenshots by meaning or keywords - the general image search. Semantic and keyword matching run together and merge. Returns each image with its caption, tags, image and page URLs and a ready citation (source, licence). Use the key words only: a long descriptive phrase matches fewer images strongly. Every image carries rights (source publisher/creator/community, licence, attribution line, commercial_use yes/no/unknown - unknown is never a guess), and match_quality says whether the search found any strong match (strong / closest_only / none). When corrected_query is present, the results are for that corrected wording, not the words as typed - say so to the user. Results exclude images classified controversial unless include_controversial is true.

Find a meme or reaction image for an idea, a mood or a sentence, ready to paste with its citation. Use it when writing a post, an article or a reply and you need a meme to make the point. Returns only images tagged as memes (the gallery's Memes filter), so it may come back empty when no meme fits. To search every kind of image, use search_images. When corrected_query is present, the results are for that corrected wording, not the words as typed - say so to the user. Results exclude images classified controversial unless include_controversial is true.

Get the receipts for one image by ID - where it came from (who added it and when, the credit printed on it, its publisher and versions), what it claims about the world (each checkable claim restated neutrally, with the source the image itself names), and what the data says (charts from the original publishers - Our World in Data, Eurostat, USDA ERS, EEA - that measure the same thing, with period, unit and a link). Use it when the user asks whether an image is true, where it comes from or what backs it. No verdict: the claims are the image's own, and an empty data list means MemeStack holds no publisher chart for that measure. The first ask for an image reads it and can take several seconds.

Find a meme, chart or screenshot on MemeStack from a copy of the image - to name a meme or find where an image came from: give an HTTPS image URL, or a `data:image/...;base64,...` URL for bytes pasted into the conversation, and get the same or visually similar images with their source and citation. Matches by perceptual hash (dHash hamming distance), so it finds copies and edits, not images about the same topic. Heavily rate-limited (10/min/IP) since each call hashes the input. Results exclude images classified controversial unless include_controversial is true.

Find images that look like a MemeStack image, by its ID - the same meme, chart or screenshot, its crops and edits (perceptual-hash matching), each with a ready citation. For an image that is not on MemeStack, use reverse_image_search. Results exclude images classified controversial unless include_controversial is true.

Submit an image to the MemeStack review queue. Paid tool - $0.10 per submission, paid in sats over Lightning at the current BTC/USD rate or in USDC on Base via the agent payment rails (x402/L402); no account or API key. IMPORTANT: there is NO refund if moderation rejects the submission (duplicate, nsfw/spam, low quality) - the fee is the anti-spam mechanism. Accepted images await human review before publication. Accepts an HTTPS image URL or a data:image/...;base64 URL (JPEG/PNG/WebP, max 8MB). Optional caption/tags/attribution are hints shown to the AI tagger and moderators - never published verbatim. If the call fails or times out, retry with the SAME payment proof to resume - one payment delivers at most one submission and is never charged twice (up to 10 delivery attempts per payment; a corrected image_url is allowed while no submission has been created yet).

Find screenshots, memes and other images by the exact words printed in them (OCR text): a quote, a caption, a sign, a watermark. Matches only the text in the image (the text_in_image field), not captions or tags, and returns strong matches only, each with a ready citation. Results exclude images classified controversial unless include_controversial is true.