What Scoutpost Is

Scoutpost, formerly coJournalist, is an AI-powered research assistant for journalists. It monitors websites, local news sources, social media, and government documents on your behalf, and delivers findings via email and dashboard.

It is not a news source. All content displayed in the dashboard, delivered via email, and generated in export drafts is produced or processed by artificial intelligence. This includes extracted facts, article summaries, translated titles, filtered search results, and draft articles.

AI Content — Read This

AI-generated content may contain errors, omissions, hallucinations, or misattributions. Scoutpost does not verify the accuracy of extracted information. Specifically:

• Extracted facts may be incomplete, taken out of context, or wrong
• The AI filter may exclude relevant articles from your results
• Translated titles may distort the original meaning
• Export drafts may contain fabricated connecting details or inferred causation
• Source reliability varies — community and independent sources have no editorial oversight

You are solely responsible for independently verifying all information before publication. Treat Scoutpost output as unverified leads, not confirmed reporting. We are not liable for editorial decisions made based on AI-generated content.

Your Responsibilities

When you use Page Scout, Civic Scout, or Social Scout, you are responsible for ensuring that use complies with applicable target website and platform terms. You must comply with the verified source and service constraints that apply.

Social Scout monitors publicly accessible personal profiles only. If Scoutpost cannot confirm that an account is a public profile, it warns you and continues; you remain responsible for confirming the account is public.

Acceptable Use

Scoutpost is for investigative work with a public-service purpose. It must not be used for personal surveillance, harassment, coercive employee monitoring, stalking, or retaliation.

The following uses are prohibited:

• AUP-PRIVATE-ACCESS — Unauthorized access to private content or circumvention of access controls is prohibited.
• AUP-STALKING-HARASSMENT-DOXXING — Using Scoutpost for stalking, harassment, or doxxing is prohibited.
• AUP-COERCIVE-DISCRIMINATORY-SURVEILLANCE — Coercive or discriminatory surveillance, including coercive employee monitoring, is prohibited.
• AUP-PROMPT-INJECTION — Prompt injection attempts to override Scoutpost instructions are prohibited.
• AUP-RATE-LIMIT-ABUSE — Rate-limit abuse of Scoutpost or source services is prohibited.
• AUP-MISINFORMATION — Using Scoutpost to generate or amplify misinformation is prohibited.

Monitoring a publicly accessible personal profile for a legitimate public-interest purpose is not by itself a violation. Attempting to access private content is prohibited.

Potential violations receive operator review; model output alone is never proof. High-confidence findings require corroborating evidence and a cited published rule. An operator may terminate an account after review of relevant logs and may share those logs when appropriate and lawful. Scoutpost takes no enforcement action without that review.

The stable rule identifiers, evidence requirements, and uncertainty treatment are published in the machine-readable acceptable-use policy.

No Warranty

Scoutpost is provided "as is" without warranty of any kind. We do not guarantee completeness of search results, accuracy of translations, availability of social media scraping (platform changes may cause failures), or continuity of any feature or data source.

AI filtering may exclude relevant articles. The system searches a subset of available sources. Important stories may be missed. Scoutpost supplements but does not replace your own monitoring and sourcing.

Intellectual Property

You own the content you create using Scoutpost's export feature. AI-generated drafts are tools for your editorial process — we claim no rights over your published work.

Account & Deletion

Hosted Scoutpost users can permanently delete their Scoutpost account and associated data from Preferences > Account. A Scoutpost Pro or Team plan must first be cancelled or downgraded through MuckRock; deleting Scoutpost does not cancel or delete the MuckRock account. If self-service deletion is unavailable, contact support. Self-hosted deployments use the authentication and deletion process configured by the newsroom.

What We Collect

• User ID — an opaque Supabase user identifier
• Scout configurations — URLs, locations, topics, criteria, schedules you set up
• Location data — city, country, and coordinates for geo-targeted scouts
• Extracted facts — AI-extracted statements, source URLs, dates, entities
• Social post content — public post text and metadata collected by Social Scouts
• Execution history — when scouts ran and what they found
• Usage records — credit consumption for billing
• Abuse-review records — policy rule, deterministic evidence, confidence, model rationale, notification state, and operator disposition for potentially abusive scout configurations
• Subscription claim metadata — for hosted Indicator claims, a Beehiiv subscription ID and keyed fingerprint used to verify entitlement without storing the submitted subscription email

What We Don't Collect

• Passwords in application storage (authentication is handled by Supabase Auth)
• Analytics, behavioral tracking, or advertising data
• A record of your general browsing activity or behavior across websites
• Indicator subscription emails in application tables after a claim link is sent

Scoutpost Page Scout for Chrome

This section covers the Chrome extension and its connection to hosted Scoutpost. Its purpose is to let you test a public page and create a scheduled Page Scout from that page. The privacy protections above and below also apply; the details here explain the extension's specific data handling, including data kept only in your browser.

• Current page and your inputs. When you open the popup, the extension reads the active tab's URL and title to prepare an editable draft. The title becomes the suggested Scout name. It does not read your full browsing history, watch tabs in the background, extract the page body from your browser, or access browser cookies. It uses no content scripts. The selected URL and title are nevertheless browsing-activity and website-content data, not an exception to disclosure.
• Testing and monitoring. Clicking Test sends the URL and any criteria to Scoutpost. Confirming Create also sends the Scout name, topic labels, schedule and a request identifier. Scoutpost fetches the public page on its servers, rather than using your logged-in browser session, and processes the returned website content to test accessibility, evaluate criteria and establish a monitoring baseline. A created Scout continues fetching on its schedule and can produce findings and email notifications. Review URLs and criteria before sending them; they may themselves contain personal or sensitive information.
• Account and authentication. Hosted sign-in uses MuckRock and Supabase. The extension receives your Scoutpost account ID and email to identify the connected account, plus opaque access and refresh tokens for the permissions you approve. It does not receive your account password. Scoutpost stores hashes of the extension's authorization codes and tokens, along with permission grants, client/account identifiers, consent and expiry information, and last-use timestamps.
• Local session data. Tokens, account details, sign-in state, drafts, test results and creation-recovery details use Chrome's extension session storage, restricted to trusted extension contexts, not synced storage or persistent extension storage. They survive closing the popup but not the end of the browser session. Discarding a draft removes its local draft and preview. A successful Disconnect removes local credentials and clears the preview and success display, but preserves the draft and unresolved creation details so you can recover the outcome after reconnecting. Dismissing a request does not cancel server work.
• Network and security data. Our hosting and network providers process your IP address and request metadata when you connect. For authorization rate limits, Scoutpost receives the verified client IP through its proxy and uses a hash of that IP in a client-specific rate-limit record. An IP address can indicate an approximate location; the extension does not request precise device geolocation or use IP data for behavioral tracking. Account/grant rate limits and short-lived operation records also help prevent abuse and duplicate work.

Recipients and processors. Buried Signals operates Scoutpost. Cloudflare handles the dedicated authorization ingress; Render hosts the website/API proxy and server-side crawling infrastructure; Supabase handles authentication, Edge Functions and stored application data. These services process the requests and data needed for those roles, including connection metadata. MuckRock handles hosted sign-in, and Beehiiv supports Indicator entitlement checks as described below. Public page fetching uses our Crawl4AI service, with Firecrawl as a fallback for supported access failures; the target website receives the server-side fetch request. Page content and criteria needed for AI evaluation or extraction pass through OpenRouter to Google Vertex. Resend receives the email address and notification content when a Scout sends email. Not every processor listed in the general policy is used by every Page Scout.

Security and Limited Use. Scoutpost Page Scout's use and transfer of information received through the Chrome extension adheres to the Chrome Web Store User Data Policy, including its Limited Use requirements. We use this data only to provide and maintain the disclosed Page Scout features and secure their operation. Transfers are limited to what is necessary for those features (including the processors described here), security and abuse investigation, or compliance with applicable law. We do not sell extension user data, use or transfer it for advertising, personalized or interest-based ads, retargeting, creditworthiness assessments or lending decisions, or use it for unrelated profiling.

Production extension requests to Scoutpost use HTTPS. Access is constrained by explicit consent, limited permissions, expiring tokens, rotating refresh tokens and server-side account checks. Human access to extension user data is limited to your explicit consent for specific data (for example, support), necessary security or abuse investigations, or compliance with applicable law. These protections do not mean that data is end-to-end encrypted or that operators can never access it.

Extension-specific retention. Expiry stops authorization; it is not the same as deleting a database record. The scheduled cleanup runs hourly:

• Authorization requests and single-use codes expire 10 minutes after the request starts. Their records become eligible for cleanup one day after expiry, or earlier when their associated grant or account is deleted.
• Access tokens expire after one hour; expired access-token hashes are removed by cleanup. Refresh tokens expire after 30 days and are rotated on use. A grant expires 30 days after consent or its latest successful refresh. Used refresh-token hashes are retained with the grant to detect replay.
• Grant records are removed by cleanup when expired, or when revocation is more than 30 days old, whichever comes first. Deleting a grant also removes its remaining token and associated authorization records. Revocation blocks further authorized calls without waiting for cleanup.
• Creation receipts contain account/client/request identifiers, a hash of the submitted configuration, a Scout identifier, and the outcome, which can include Scout configuration or an error. They remain while the Scout exists. If no Scout exists, including after a failed creation or Scout deletion, cleanup records that absence and removes the receipt after more than 30 days of recorded absence. They are not deleted merely because a connection is revoked or its grant expires.
• Rate-limit records become eligible for cleanup one day after their rate-limit window starts. Operation leases expire after 10 minutes and are normally removed when the operation finishes; expired leases are also cleaned up.
• Account deletion removes account-linked grants, authorization records, tokens, operation leases and creation receipts. Scout configurations, fetched content, findings and notifications are separate from the connection: the general retention rules below apply. Ordinary raw page captures have a 30-day retention period, but current comparison baselines may be retained for the Scout's lifetime.

Your controls and rights. Disconnect in the extension or revoke the connection in Scoutpost Preferences > Connected integrations. Uninstalling or ending the browser session clears local extension data but does not itself revoke the server-side grant. Disconnecting, revoking, uninstalling or discarding a draft does not stop or delete existing Scouts; pause or delete them in Scoutpost. You can request access, correction, deletion or a machine-readable copy of your data using the rights described below. For extension privacy questions or data requests, including when you cannot sign in, email tom@buriedsignals.com. Account deletion follows the Account & Deletion process above; connection revocation is not an account-deletion request.

Scoutpost for ChatGPT

This section covers the Scoutpost plugin for ChatGPT and Codex, which connects those apps to your hosted Scoutpost account through the Scoutpost MCP server. Its purpose is to let you create and manage scouts and work with their findings from a conversation. The rest of this policy also applies. OpenAI operates ChatGPT and Codex; OpenAI's own privacy policy, not this one, governs how OpenAI handles your conversations and the data it receives.

• What ChatGPT sends to Scoutpost. The inputs of the tools it calls for your request: for example scout names, URLs, locations, topics, criteria, schedules, public social profile handles and council websites; search queries; project names and notes; text or URLs you ask it to ingest; and verification notes on findings. The plugin does not request your conversation history, browser cookies, logged-in sessions, precise location or account password. Scoutpost fetches public sources on its own servers. Scouts keep running on their schedule, can send email notifications and use credits. Review inputs before confirming; URLs, criteria and notes may themselves contain personal or sensitive information.
• What Scoutpost returns to ChatGPT. Tool results go to OpenAI and appear in your conversation. Depending on the request they include your scouts and their settings and status; findings (information units) with their source URLs, extracted facts and verification state; Civic items and run summaries; evidence snapshot lists and short-lived download links; projects, notes and entity names; and the identifiers needed to act on these records. Findings and entities can name people who appear in the public sources your scouts monitor.
• Account and authentication. You connect through ChatGPT's sign-in flow, which registers ChatGPT as a client of the Scoutpost MCP server and opens hosted Scoutpost sign-in (MuckRock and Supabase). ChatGPT then receives a Scoutpost session access token, valid for one hour, and a refresh token for your account. When it refreshes the session, it also receives your Scoutpost account ID, email address and basic account profile details. It never receives your password, and the plugin never asks for passwords, API keys or session tokens in conversation. Scoutpost stores the client registration (client name, redirect addresses and last use), which identifies the ChatGPT connection rather than you, and briefly stores single-use authorization codes.
• Network and security data. Our hosting and network providers process your IP address and request metadata when you sign in. Tool calls come from OpenAI's servers, so their network metadata identifies OpenAI rather than your device.

Purposes and recipients. We use this data only to provide the Scoutpost features you request through the plugin, secure them, prevent abuse and comply with applicable law. OpenAI receives the tool results and tokens described above. The other recipients and processors are those listed under Third-Party Processors below. We do not sell this data or use or transfer it for advertising or unrelated profiling.

Retention. Authorization codes expire after 10 minutes and are deleted by a daily cleanup once used or expired. Client registrations are kept so returning connections are recognized. Scouts, findings, projects, notes and notifications created through the plugin follow the general retention rules below, the same as records created in the app. Copies of tool results and tokens held by OpenAI are subject to OpenAI's retention.

Your controls and rights. Disconnect Scoutpost in ChatGPT or Codex to remove their copy of your tokens. To have Scoutpost revoke the connection's session on its side, email support; deleting your account also ends it. Disconnecting does not stop or delete existing scouts; pause or delete them in Scoutpost. Deleting a ChatGPT conversation does not delete data held by Scoutpost. You can request access, correction, deletion or a machine-readable copy of your data using the rights described below. For plugin privacy questions or data requests, email tom@buriedsignals.com.

Third-Party Processors

Your data is processed by:

AI requests therefore follow Scoutpost → OpenRouter → Google Vertex. OpenRouter remains a processor; routing through it does not reduce the processor chain. Scoutpost restricts each request to Google Vertex, requests zero-data-retention routing, denies provider data collection, and disables OpenRouter response caching. Account-level logging and data-sharing settings provide additional protection.

PDF text extraction normally runs locally with Poppler. Only low-yield or scanned PDFs are sent through the AI route for native Google PDF processing; Scoutpost forces the native PDF engine and does not use OpenRouter's Mistral or Cloudflare parsing options.

Data Retention

Your Rights

You have the right to:

• Access — request a copy of all data associated with your account
• Deletion — request permanent deletion of all your data
• Rectification — correct any inaccurate data
• Portability — receive your data in a machine-readable format

You can download a machine-readable (JSON) copy of the data stored for your account yourself from Preferences > Your data, up to three times per hour. The file covers your account, settings, credit usage, scouts and their runs, extracted information, entities, projects, civic trackers and connected tools. It leaves out embedding vectors derived from your text, security secrets such as key hashes and tokens, internal processing records, and the bodies of captured pages and posts, which it lists by URL and time instead. To obtain anything the file does not include, or to exercise another right, contact us as described below.

Data is processed in the region configured by your newsroom or hosting provider. EU users may lodge a complaint with their local data protection authority.

Cookies

Scoutpost uses the authentication storage mechanism configured for your deployment. Depending on your setup, this may include secure cookies or browser-managed session storage.

We do not use analytics, tracking, or advertising cookies.

Contact

For data requests or to exercise your rights under GDPR: click the help icon inside the app — it opens a private channel to the maintainer. For general questions, open an issue on GitHub.

Support. For help with hosted Scoutpost, the Chrome extension or the ChatGPT plugin, including when you cannot sign in, email tom@buriedsignals.com.

Buried Signals · Zurich, Switzerland