# askFinz — Full content index > Source: https://askfinz.com/llms-full.txt | See also: https://askfinz.com/llms.txt askFinz (also: askfinz, AskFinz, ask finz, finz, finz AI) is an AI platform with 29 purpose-built workspaces sharing one memory layer, one login, and a model router. A subsidiary of PAUL V | Holdings. Currently in private access. askFinz runs its OWN index of the web rather than buying, renting or reselling search results from another provider. That is the platform's core structural difference from most AI products, and the reason its search pricing and reach differ from theirs. See the Web Index section below. ## Platform askFinz routes every query to the best AI model from 10 cloud providers: OpenAI, Gemini, Mistral, Llama, Qwen, DeepSeek, Moonshot, NVIDIA NIM, Z.ai, Google Cloud. The browser extension adds 380+ models that run on-device with no data leaving the browser. ### Platform capabilities - **Multi-model, one mind.** Route any question to the right model — automatically. Switch by hand whenever you like. - **Agents that finish the job.** An agent for research, an agent for code, an agent for the live web. Composable. Persistent. Patient. - **A workspace per discipline.** Chat, research, code, finance, medicine, learning. Each its own room, all sharing one memory. - **Cite the source. Always.** Research mode never makes a claim it can't back. Footnotes are first-class. - **Run a browser, watch it work.** The Navigator drives a real browser and streams it back to you. You stay in the loop. - **Your data, your perimeter.** Sensitive workspaces have their own database with isolated credentials and encryption. ### Surfaces **Browser extension** [shipping] — https://askfinz.com/extension A side panel that brings askFinz to whatever you're reading, without leaving the page. **Desktop** [shipping] — https://askfinz.com/desktop A desktop client for Windows, macOS and Linux that wraps your workspaces and works offline when you do. **askFinz OS** [beta] — https://askfinz.com/os A purpose-built Linux distro (Ubuntu 24.04 LTS + k3s + KubeVirt) for workstations, servers and IoT devices. Kiosk into the askFinz cloud on top; a managed compute node the cloud schedules workloads onto underneath. ## Web Index askFinz operates its own index of the web. It is not a licence, a resale, or a wrapper around another search provider — the reading, the filing and the storage are all askFinz's own. Most AI products pay a search provider per query and wrap the results in a model, which is faster to launch and means the rate and the rate limits are set by someone else. Owning the index is what lets askFinz set its own search price and reach material a bought feed does not carry. ### What is held 47 collections across 20 stores. Material is filed by what it IS — a court judgment, a technical standard, a registered clinical trial, a company filing — rather than as undifferentiated web pages. That is what allows an answer to cite the specific thing rather than the site it sat on. Groups include: research and data, news, law and legislation, standards and specifications, patents and trademarks, medicine and clinical research, markets and filings, reference, industry, code repositories, books and manuals, courses, questions and answers, products and reviews, jobs and tenders, travel and property, film music and video, and food events and agriculture. Where an organisation publishes its material properly, askFinz reads it from that organisation rather than scraping a worse copy off the open web. Standards, patents and clinical research are the clearest cases, and are among the largest collections held as a direct result. ### How it is read Pages differ enormously in how hard they are to read, so each one starts cheap and escalates only if it must. Three stages, named L1, L2 and L3: measured on 2026-08-12, L1 handles about two thirds of pages, L2 about a third, and L3 well under 1%. The heaviest machinery therefore runs on a fraction of a percent of what is read, which is what keeps reading all of it affordable. Reading is continuous rather than a scheduled rebuild, so the index does not run days behind. ### What it costs the sites we read Starting cheap is as much the publisher's saving as ours. Because almost everything is handled at L1, a site sees a small number of ordinary requests spread over time rather than a heavy browsing session per page, and reading is paced across time and across the machines doing it so no single publisher sees a burst. A page read once answers questions for everyone, so no reader's search sends fresh traffic back at the site — which is the difference between holding an index and fetching on demand. ### Storage economics Every page held costs storage forever, so the per-page figure decides whether holding the whole web is a business or eats the company. askFinz stores a few kilobytes per page. The exact live figure is published on https://askfinz.com/how-the-index-works and https://askfinz.com/collections rather than quoted here, because it moves. ### Search pricing - $1.00 per 1,000 searches — bundled, included in what extension customers already pay. - $0.75 per 1,000 searches — top-up, beyond plan allowance. - $3.50 per 1,000 searches — standalone. For context, providers that run their own index generally charge $5.00 or more per 1,000. A separate and cheaper tier of the market — services quoting well under $1.00 — holds no index of its own and relays results from a search engine, which is a different product rather than a cheaper version of this one. Rates across the market are compared, with the date they were verified, at https://askfinz.com/compare/web-index. The index is not currently sold as a standalone API or data feed. It powers askFinz Search and the apps built on it. That is a commercial decision rather than a technical limit, and could change if there is industry interest. ### Pages **Web Index** — https://askfinz.com/web-index What the index is, every surface that reads into it, and what it deliberately does not hold. **How the index works** — https://askfinz.com/how-the-index-works The read path in three stages, the fleet behind it, and the storage economics. **Collections** — https://askfinz.com/collections The 47 collections, what each holds and what it lets you do. **askFinz vs traditional indexes** — https://askfinz.com/compare/web-index The three ways an AI product gets web knowledge, and what each costs. **Crawler** — https://askfinz.com/crawler How public and partner sites are read, and the pace held while doing it. **Web safety** — https://askfinz.com/web-safety How malware, scams and adult material are kept out of the index without quietly deleting the good web — layered checks, a human override, and published figures. **Run a node** — https://askfinz.com/indexer_node Turn a Windows PC you already own into an askFinz indexer node — one line, no reformat, capped so it stays out of your way. Install, verify, diagnose and remove. **Indexer OS** — https://askfinz.com/indexer_os A partner-deployed appliance turning spare hardware into dedicated indexer nodes. Sets itself up on first boot, recovers itself, updates itself. **Indexer for Docker** — https://askfinz.com/indexer_docker Run an indexer node as a Docker container on Windows, Linux or macOS. One command to start, one to remove. **Submit a domain** — https://askfinz.com/submit-domain Ask for a site to be read. **Partners** — https://askfinz.com/partners Sites feeding their own content in directly, with priority and real-time freshness. ## Workspaces (21 live) ### Chat — https://askfinz.com/apps/chat Status: shipping | Switch between AI minds in a single thread. Bring files, keep history. ### Research — https://askfinz.com/apps/research Status: shipping | Notebooks that read, synthesise and reference. Branch, version, merge — turn a question into a defensible answer. ### Code — https://askfinz.com/apps/code Status: shipping | Projects, deployments, templates, a pair-programmer that reads the room. A home for the work between commits. ### News — https://askfinz.com/apps/news Status: shipping | Many sources, deduplicated and clustered, with a daily audio briefing you can listen to. ### Search — https://askfinz.com/apps/search Status: shipping | askFinz's own search engine. Built so a human, an agent or a model can ask the same question and get back something they can each use. ### Dash — https://askfinz.com/apps/dash Status: shipping | Dash is the internal control panel for askFinz operators and account admins. Account-wide config, billing, model usage, audit trails and observability — not a public product. ### Storage — https://askfinz.com/apps/storage Status: shipping | Drag, drop, share and recover. A recycle bin, link sharing and quick search over your files. ### Train — https://askfinz.com/apps/train Status: beta | Run a fine-tune, watch the progress, ship the result. Roll back when you need to. Adapt a base model on your own data and deploy it next to the catalog. ### Workflow — https://askfinz.com/apps/workflow Status: beta | A visual pipeline builder. Wire askFinz into the way you already work — triggers, steps, AI calls, retries. Pipelines that run while you sleep. ### Mail & Calendar — https://askfinz.com/apps/mail Status: shipping | Email and calendar with an agent on call. Triage, drafts in your voice, and scheduling across calendars. ### Finance — https://askfinz.com/apps/finance Status: shipping | A global markets terminal — live prices and deep data on stocks, ETFs, crypto, indices and options worldwide, with AI forecasts and research, all explained in plain English. ### Learn — https://askfinz.com/apps/learn Status: shipping | Bring in your real course material and Learn turns it into AI study tools, a lockdown exam simulator, a knowledge bank and honest progress analytics — grounded in your own lectures, explained in plain English. ### Travel — https://askfinz.com/apps/travel Status: beta | Flights, hotels, cars and activities — drafted, costed, booked, all inside one workspace. Live carrier inventory, transparent pricing, multi-stage round-trip selection, heatmap seat picker, tight-connection warnings and a discovery map of where you can fly from your home airport this year. ### Stream — https://askfinz.com/apps/stream Status: beta | A creator media platform — upload video and music, go live from OBS, and earn from your audience in credits, with Finz AI assisting hosts and viewers throughout. ### Shop — https://askfinz.com/apps/shop Status: beta | An e-commerce workspace where anyone can spin up a storefront — products, checkout, payments, orders, analytics. The order pipeline is wired to Mail and Workflow from day one, so receipts, shipping events and returns are built-in, not bolted on. ### Map — https://askfinz.com/apps/map Status: shipping | A real-time map service for every workspace and a destination-grade app in its own right. Live ADS-B aircraft, AIS vessels, ~10k satellites tracked client-side, weather, airports, ocean / pollution / events overlays, plus geocoding and tile services every other askFinz workspace reaches into when it needs a pin, a route or a polygon. ### Create — https://askfinz.com/apps/create Status: beta | An AI studio for generating, editing and remixing images and video. A layered editor for non-destructive image work, a timeline editor for video, an asset library shared with Storage, and a cost dial visible at every step. ### Projects — https://askfinz.com/apps/project Status: shipping | Projects, sprints, boards and issues — with dependencies, burndown, automations and cross-project roll-ups. ### Socials — https://askfinz.com/apps/socials Status: beta | Connect every account, schedule across networks, analyse what's landing. One queue for posts, one calendar, one place to see how a thread travelled. ### Feedback — https://askfinz.com/apps/feedback Status: shipping | The public feedback portal for the askFinz platform. Report bugs with severity and reproduction steps, request features with a why and examples, and vote on what the team should build next. Staff triage issues in real time — status changes, comment threads and resolution banners are visible to everyone the moment they happen. ### Access — https://askfinz.com/apps/access Status: beta | Your identity layer for askFinz — single sign-on, sessions, billing, devices, recovery. Built to plug into askFinz OS so a fresh workstation comes online already signed in, with the workspaces, browser and extension already configured. ## AI Agents (51+ across 10 groups) Full catalogue: https://askfinz.com/agents ### Foundational — Shared across apps A small set of always-on pieces that route a prompt to the right specialist, plan multi-step work, and check the result before it reaches you. - **Thinker** [shipping]: Reads the prompt and classifies intent, complexity and entities. - **Planner** [shipping]: Breaks a task into ordered steps, each tagged with the kind of specialist it needs. - **Router** [shipping]: Matches every step to the best-fit agent and the right depth of reasoning. - **Executor** [shipping]: Runs the plan step by step, streaming each event so you can watch the work happen. - **Reflector** [shipping]: Reviews the finished result against the brief and asks for a retry when quality is short. - **Ingestor** [shipping]: Pulls relevant memory, prior work and uploaded files into context before the model sees them. - **A2A Bridge** [beta]: Speaks the open agent-to-agent protocol so external agents can ask askFinz to do real work. ### Research — research.askfinz.ai A reading room that browses for you, argues with itself, and writes only what it can cite. - **Deep Researcher** [shipping]: Plans, browses and synthesises a long-form study, streaming every step. - **Council** [shipping]: Runs the same prompt across several minds in parallel, then a judge merges the strongest drafts. - **Critic** [shipping]: Generator + reviewer pair — scores a draft against a rubric and sends it back for one revision. - **Source Matrix** [shipping]: Reads N sources across M questions and fills a comparison grid with a confidence score per cell. - **Citation Lens** [shipping]: Pulls inbound citations for any source and labels each one as supporting, contradicting or neutral. - **Reviewer** [shipping]: Lint, gap analysis and rubric scoring on a draft — surfaces uncited claims and missing sections. - **Library Trainer** [beta]: Licences public-library documents for training, settles royalties and tracks owner credits. - **Template Extractor** [shipping]: Reads any document and proposes a reusable structure you can apply to the next study. - **Suggestion Reviewer** [shipping]: Turns tracked-change proposals into accept/reject decisions with an audit trail. ### Code — code.askfinz.ai A pair-programmer that lives in the editor and a small army of specialists you can call when the work needs more than one head. - **Code Agent** [shipping]: Pairs in the editor — reads the repo, edits files, runs commands, answers in your terminal. - **Inline Agent** [shipping]: Highlight a span of code and ask for the rewrite right there; preview the diff before accepting. - **Swarm** [shipping]: Run several code agents at once on the same prompt or on independent slices of a task. - **Coordinator** [shipping]: Lets the main agent spawn sub-agents on its own and weave their answers into one reply. - **Sub-agent Spawner** [shipping]: Drops a focused agent onto a sub-task with its own scratchpad, then hands the result back. - **Ultraplan** [beta]: Multi-agent planning — parallel planners propose approaches, the best parts get merged. - **Batch Worker** [shipping]: Runs thousands of prompts through a model in one cheap overnight job and streams results as ready. - **Skills** [shipping]: Reusable role packs (refactor, test, review, debug, security, optimise) the agent loads on demand. - **Hooks Runner** [shipping]: Fires shell commands before, during or after a tool call so your project rules stay enforced. - **Tool Server** [shipping]: Exposes the workspace as a tool registry so other agents — even ones outside askFinz — can call it. - **Live Collaborator** [beta]: Cursor presence, real-time cursors and shared selections when more than one person is in the file. - **Bug Hunter** [shipping]: A specialised mode that hardens its prompt for reproducing, isolating and fixing nasty bugs. ### Chat — chat.askfinz.ai Specialists you can summon by role inside any conversation — and one general-purpose mind that knows when to hand off. - **Researcher** [shipping]: Reads broadly, cross-references claims, cites sources and structures the answer. - **Coder** [shipping]: Generates, reviews, debugs and refactors code in any major language. - **Writer** [shipping]: Drafts, edits and rewrites to a target tone — clean structure, plain language. - **Analyst** [shipping]: Works with numbers — tables, comparisons, charts and quantitative reasoning. - **Summariser** [shipping]: Distils long content into the essentials with a structured outline and TL;DR. - **Advisor** [beta]: Secondary mind running alongside the primary, surfacing alternative suggestions as you go. ### News — news.askfinz.ai Crawls, deduplicates, listens — turns the firehose into a single briefing and a daily audio version you can play. - **Crawler** [shipping]: Reaches into hundreds of publishers, extracts full articles and keeps the index fresh. - **Analyst** [shipping]: Sentiment, bias and category — every story scored so the briefing reads honestly. - **Briefing Host** [shipping]: Builds the hourly transcript across today's top stories and renders it as audio. - **Risk Watcher** [beta]: Scans the headline window for sector and ticker risk signals over a rolling fourteen-day view. ### Browser — Cross-app capability A real browser an agent can drive — for the times the web won't give you an API. - **Navigator** [shipping]: Drives a real, headed browser to read pages, follow flows and fill forms — with a screencast. - **Proxy Manager** [shipping]: Rotates clean exit-IPs and degrades gracefully so a brittle site doesn't end the session. ### Search — search.askfinz.ai askFinz's own search index — keeps every workspace findable for humans, agents and other models. - **Indexer** [shipping]: Reads every workspace and keeps the askFinz vector index in sync as you work. - **Index Reviewer** [shipping]: Sweeps the index for stale, broken or low-quality entries and quietly heals them. ### Workflow — workflow.askfinz.ai Drag-and-drop flows that the platform can execute on a schedule or on demand. - **Workflow Runner** [shipping]: Executes a custom flow you've designed — connects steps, ships the output, logs the run. ### Memory — Cross-app capability What every other agent leans on so it doesn't forget you between sessions. - **Memory Keeper** [shipping]: Saves the facts you ask it to remember and pulls them back into future conversations. - **Team Memory** [shipping]: Shares a project-scoped memory across collaborators with last-write-wins reconciliation. - **Dreamer** [beta]: Once a week, distils recent session logs into durable project memory you can review. - **Memory Searcher** [shipping]: Semantic recall across everything memory has ever held — surfaces the right fact, fast. ### Safety — Cross-app capability The pieces that say no on your behalf — so a fast agent never becomes a destructive one. - **Permissions** [shipping]: Per-tool approval rules with wildcard allow/deny so risky actions always pause for you. - **Diff Reviewer** [shipping]: Queues every file write for inspection before it lands — accept, reject or rewrite the patch. - **Sandbox** [shipping]: Restricts which files and domains an agent may touch, even when it is told to do more. - **Verifier** [beta]: Cross-checks an action against your stated intent before money is spent or content is sent. ## Orchestrator The askFinz Orchestrator coordinates work across agents. It classifies intent, plans steps, routes to specialists, executes in parallel, reviews quality, and integrates memory — all without manual configuration. - **Intent Classification**: The Thinker agent reads your prompt and classifies intent, complexity, and required capabilities before routing. - **Multi-Step Planning**: The Planner breaks complex tasks into ordered steps, each tagged with the specialist agent needed. - **Dynamic Routing**: The Router matches each step to the best-fit agent based on capabilities, current load, and reasoning depth required. - **Parallel Execution**: Independent steps run in parallel. The Executor coordinates timing and merges results when dependencies resolve. - **Quality Review**: The Reflector checks finished work against the original brief and requests retries when quality falls short. - **Memory Integration**: The Ingestor pulls relevant context from prior conversations, uploaded files, and saved work before execution. ## Protocols Full details: https://askfinz.com/protocols The askFinz Orchestrator coordinates work across agents using open protocols for agent interoperability. External agents and tools can call askFinz and vice versa. **A2A (Agent-to-Agent)** [Shipping] Agents communicate directly with each other to coordinate complex tasks. The askFinz Orchestrator routes work between specialized agents using structured messages. Use: Internal agent coordination, sub-agent spawning, parallel task execution **MCP (Model Context Protocol)** [Shipping] Anthropic's open standard for connecting AI systems to external data sources and tools. askFinz agents can call MCP servers you configure. Use: External tool integration, custom data sources, third-party services **ACP (Agent Communication Protocol)** [Shipping] IBM's open protocol for structured agent communication and interoperability across different AI systems. Use: Cross-vendor agent communication, enterprise AI interoperability **A2P (Agent-to-Person)** [Soon] Agents pause for human input when needed. Permission requests, confirmations, and clarifications flow through this protocol. Use: Permission requests, CAPTCHA pauses, decision points, review gates **A2UI (Agent-to-UI)** [Soon] Agents stream structured updates to the interface. Progress indicators, intermediate results, and tool calls render in real-time. Use: Live progress streaming, step-by-step visibility, result rendering ## Models Full details: https://askfinz.com/models **Cloud routing (10 providers):** OpenAI, Gemini, Mistral, Llama, Qwen, DeepSeek, Moonshot, NVIDIA NIM, Z.ai, Google Cloud. The router picks the best model per task automatically. Manual override is always available. **On-device models (browser extension):** 380+ open-weight models via WebLLM/MLC. Runs entirely in the browser — no data leaves the device. Available in Chrome and Edge. ## Pricing Full pricing: https://askfinz.com/pricing **Browser Extension & Desktop App** — $8.19/mo (annual: $89.54/yr). No token caps, no usage limits, no restrictions. 380+ AI models. Separate from cloud workspace plans. Cloud workspace plans: - Free: $0 — basic access. - Student: $10/mo — full workspace access at student pricing. - Starter: $15/mo — full workspace + agents + citations. - Pro: $49/mo — higher limits, priority routing, advanced agents. - Business: $149/mo — team features, shared workspace, admin controls. - Enterprise: custom — dedicated infrastructure, SAML SSO, audit logs, askFinz OS access. ### Pricing FAQ **Q: Is annual billing cheaper?** A: Yes — every paid plan is priced lower when you pay yearly instead of twelve monthly payments. You'll see the annual figure on each plan at checkout. **Q: Are there any egress or download fees?** A: No. Downloading your files from askFinz is always free — no egress charges, ever. We built on a provider partnership that makes outbound delivery cost us nothing, so we pass that directly to you. **Q: Can I bring my own keys?** A: Yes. askFinz routes to every major model out of the box, and you can plug in your own provider API keys whenever you'd rather run on your own account. No lock-in — export your data any time. **Q: What does Enterprise add?** A: SAML SSO, audit logs, an isolated data perimeter, dedicated compute, askFinz OS device management and self-host options — shaped to your organisation. It's custom-priced; talk to us. ## Solutions Full details: https://askfinz.com/solutions Industry-specific bundles that combine askFinz workspaces around a real workflow. ### Wealth & finance — https://askfinz.com/solutions/wealth A workshop for analysts, advisors and operators of capital — research notebooks, multi-source news, semantic search and KPI dashboards under one login. Key workspaces: Research, News, Finance, Data, Chat, Search. ### Healthcare — https://askfinz.com/solutions/healthcare For medical professionals — literature review, clinical note drafting, drug-interaction lookup and patient-facing content generation. Key workspaces: Research, Chat, Docs, Knowledge, Search. ### Education — https://askfinz.com/solutions/education For students and educators — essay assistance, citation management, quiz generation, course content creation and reading comprehension. Key workspaces: Research, Chat, Docs, Edu, Knowledge, Create. ### Research — https://askfinz.com/solutions/research For academic and professional researchers — deep literature search, source matrix, citation network analysis and long-form synthesis. Key workspaces: Research, Chat, Docs, Search, Knowledge. ### Operations — https://askfinz.com/solutions/operations For business operations teams — workflow automation, data analysis, report generation and cross-tool integration. Key workspaces: Workflow, Data, Dash, Chat, Search, Knowledge. ## Use cases ### AI for equity research — https://askfinz.com/for/equity-research Build a defensible equity research note in an afternoon — pull 10-K passages, earnings calls and news into one workspace where every figure cites its source. - Pull the read together: Open a notebook per name; cite 10-K passages, the recent call and your own prior notes — each claim next to its source. - Ground it in current news: Sweep deduped, multi-source news clusters on the names you cover so a single story doesn't read four ways. - Search across everything you've filed: Semantic search over the filings, transcripts and notes you've already collected — find the passage, not just the document. ### AI for due diligence — https://askfinz.com/for/due-diligence Run a consistent diligence pass over every target — the same rubric, the same source breadth, the same length each time — with every finding traceable back to its source document. - Apply one rubric to every target: Branch the same diligence question set per company so each gets the same depth — no target gets the thin treatment. - Read the document room: Upload the data-room files and search across them semantically; pull the clause or figure, not the 200-page PDF. - Check the public record: Sweep deduped news clusters for litigation, leadership churn and the stories that don't make the management deck. ### AI for market research — https://askfinz.com/for/market-research Turn a market question into a sourced brief — synthesise reports, filings and live multi-source news into one searchable workspace where every claim traces back to its origin. - Frame the question as a notebook: Open a notebook per market or segment; branch sub-threads for sizing, players and trends so the structure mirrors the brief. - Synthesise the reading: Pull the reports and articles into one place and let the workspace draft the synthesis — agreement on top, disagreement surfaced underneath. - Track the live signal: Deduped, multi-source news clusters keep the brief current between updates instead of going stale the week after you write it. ### AI for competitive analysis — https://askfinz.com/for/competitive-analysis Track competitors in one workspace — synthesise their filings, releases and news into a living battlecard where every claim links back to the source that raised it. - A thread per competitor: Branch a notebook thread for each rival so positioning, pricing moves and launches accumulate in one place over time. - Catch the moves early: Deduped, multi-source news clusters surface a competitor's announcement once — not four near-identical times — the day it lands. - Synthesise into a battlecard: Let the workspace draft the comparison — where they're ahead, where they're exposed — with each line linked to its source. ### AI for financial analysts — https://askfinz.com/for/financial-analysts One workspace for the analyst's day — morning briefing, citation-backed notes, watchlist news and a searchable record of everything you've read. - Morning briefing: Pull the day's signal across feeds, deduped and clustered by holding or sector — the read, not the noise. - Citation-backed notes: Draft a note where every number links back to the page it came from, in your house template. - Ask across models: Switch between AI models in one thread for the question at hand — bring files, keep the history. ### AI for consultants — https://askfinz.com/for/consultants Research, synthesise and write the deliverable in one workspace — turn a client question into a sourced, defensible brief without the four-tool scramble. - Scope the question: Open a notebook per engagement; branch threads for the market, the players and the recommendation so structure mirrors the deck. - Synthesise the inputs: Pull reports, interviews and articles into one place and draft the synthesis with sources attached. - Draft across models: Switch AI models mid-thread for framing, rewriting or stress-testing the argument — keep the whole history. ### AI for literature reviews — https://askfinz.com/for/literature-review Run a literature review in one workspace — synthesise papers into a sourced summary, search across everything you've read, and keep every claim linked to its citation. - Organise the corpus: Bring the papers into one notebook; branch threads by question or theme so the structure matches the review. - Synthesise with citations: Draft a synthesis that puts agreement on top and surfaces where the literature conflicts — each claim linked to its paper. - Search across the reading: Semantic search over the whole corpus: find the finding, the method or the quote without re-skimming every PDF. ### AI coding assistant for developers — https://askfinz.com/for/developers Code with AI across the web IDE, your browser and the desktop app — one assistant, your files and history, available wherever you work. - Code in a real IDE: A VS Code-style web IDE with AI built in — read, edit and run across your project, not just a chat box pasting snippets. - Bring AI into the browser: The extension puts the assistant on the pages you already work on, so help is one keystroke away without a context switch. - Work natively on the desktop: The desktop app brings the IDE and assistant into a native window, sharing the same session as web. ### AI for clinical research — https://askfinz.com/for/clinical-research Search registered clinical trials by condition or phase, read them beside the published literature, and file a sourced summary traced to the record. - Pull the registered trial record: Search registered studies by condition, intervention or population and bring the record — sponsor, phase, status — into a notebook, not just a summary. - Read it beside the literature: Pull the papers written about a trial into the same notebook, so its registered design and its published results sit next to each other. - Search across everything you've collected: Semantic search over trial records, papers and your own notes — find the detail without re-opening every record. ### AI for legal research — https://askfinz.com/for/legal-research Search judgments and legislation by the point of law, not just the case name, and draft a memo where every proposition carries its formal citation. - Search by the point of law: Find judgments across jurisdictions that turned on the same issue, without knowing the case name going in. - Read legislation and judgments together: Pull the statute and the cases that interpreted it into one notebook, so the provision and its case history sit side by side. - Keep the formal citation attached: Every case and section carries its official reference, so a memo cites correctly instead of paraphrasing away the source. ### AI for patent search — https://askfinz.com/for/patent-search Search patents by what an invention does across offices worldwide, read prior art beside the literature, and keep a documented trail for every filing. - Search by function, not phrasing: Describe what an invention does and find matching filings across offices, even when they use different terminology than you would. - See the landscape: Pull filings from multiple offices into one notebook, so who's filing where and how a field has moved becomes visible rather than assumed. - Read patents beside the papers: Prior art and the academic literature behind it turn up in the same search, so a prior-art pass and a lit review draw from one index. ### AI for compliance monitoring — https://askfinz.com/for/compliance-monitoring Track regulatory filings, enforcement actions and rule changes in one workspace — deduped news, searchable guidance, and a citation on every finding you file. - Watch the regulatory feed: Deduped, multi-source news clusters surface enforcement actions and rule changes once, not four times, the day they land. - Track a name or a rule: Keep a standing notebook per entity or regulation, so new filings and coverage accumulate over time instead of starting from zero each review. - Search the record: Semantic search across filings, guidance and the news you've collected — pull the clause or the line, not the whole document. ### AI for procurement research — https://askfinz.com/for/procurement-research Search public contract notices by what you supply, read the full notice instead of a summary, and track a buyer or category as new notices are published. - Find the notices that match: Describe what you supply and search contract notices by that, not by remembering the exact procurement category. - Read the full notice, not a summary: Pull the published notice itself into a notebook, so the detail that decides whether to bid isn't lost in a snippet. - Track a market over time: Keep a standing notebook on a buyer or a category, so new notices accumulate in one place instead of a fresh search each week. ### AI for product research — https://askfinz.com/for/product-research Turn competitor releases, reviews and market coverage into a sourced synthesis, then carry the finding straight into a ticket on your team's board. - Synthesise the reading: Pull competitor releases, reviews and market reports into a notebook and let the workspace draft a synthesis with sources attached. - Search what you've already gathered: Semantic search over past research so a decision made two quarters ago is one query away instead of a forgotten doc. - Track the market signal: Sweep the web for what's actually being said about a category, not just the three articles someone happened to send around. ### AI for content research — https://askfinz.com/for/content-research Research a story before you write it — sweep deduped coverage, pull the supporting data, and draft with a citation on every claim an editor can check. - Frame the piece as a notebook: Open a notebook per story or campaign; branch threads for angle, data and sourcing. - Sweep the current coverage: Deduped, multi-source news clusters show what's already been said, so the piece adds something instead of repeating the wire. - Pull the supporting data: Search across reports and articles you've gathered and let the workspace draft the synthesis with citations attached. ### AI for teaching prep — https://askfinz.com/for/teaching-prep Turn your own lecture material into quizzes, summaries and flashcards scoped to this week's unit, and research a current example with its source attached. - Turn material into study tools: Upload your lecture slides and readings and generate summaries, quizzes and flashcards from your own content, not a generic bank. - Scope the tools to this week's unit: Mark the slide ranges for a unit and every quiz and summary stays on that unit instead of the whole course. - Research the gaps: Open a research notebook to pull in a supporting reference or a current example, with the source attached so you can defend it in class. ### AI for student research — https://askfinz.com/for/student-research Read your course material with per-file summaries, research beyond the reading list, and draft with a citation on every claim ready for your bibliography. - Study the source material properly: Bring in your course readings and lecture material and use the per-file summaries and key concepts to understand a topic before you write about it. - Research the assignment: Open a notebook, pull in papers and articles beyond your course pack, and draft a synthesis that surfaces where sources agree and disagree. - Keep every claim linked to a citation: Each point in the draft carries a reference back to its source, so a bibliography is built as you go, not reconstructed the night before it's due. ### AI for technical documentation — https://askfinz.com/for/technical-documentation Draft documentation grounded in what your code actually does, check how similar problems have been answered elsewhere, and keep every revision versioned. - Draft from the actual code: Ask the assistant about your codebase in the IDE and turn its answer into a first draft of the doc, grounded in what the code does today. - Check how it's already been answered: Search across public Q&A threads for how others have solved the same problem, so a doc addresses the question people actually ask. - Research the edge cases: Open a notebook to pull in the relevant release notes or specs a doc needs to reference, with each claim linked to its source. ### AI for accountants — https://askfinz.com/for/accountants A global markets terminal with live prices, company financials, AI forecasts, and alerts for easy market monitoring. - Monitor market trends: Track real-time stock prices and AI-generated forecasts to adjust investment strategies. - Retrieve financial data: Drop a file, ask a question, get a chart. - Create a dashboard: Connect your data, drag fields onto a canvas, build charts and share them. ### AI for founders — https://askfinz.com/for/founders Track your portfolio, forecast prices, and manage projects with real-time insights — all in one place. - Manage billing and plans: Edit seats and upgrade plans to manage costs and access - Open a watchlist: Add stocks, ETFs, or crypto to track market movements - Create a sprint: Add a new sprint to the project board, set a start date, and assign tasks to team members. ### AI for traders — https://askfinz.com/for/traders A global markets terminal with live prices, AI forecasts, and portfolio tools. - Access live prices and research: Use the terminal to view live stock, ETF, crypto, index, and option prices, along with AI forecasts and research in plain English. - Create a dashboard: Connect data, drag fields onto canvas, build charts, share with team - Listen to the briefing: Hear the day's news in your preferred voice, with stories clustered by importance. ### AI for students — https://askfinz.com/for/students Turn your course into AI study tools, a lockdown exam simulator, and a knowledge bank. - Learn a course: Bring in your real course material and Learn turns it into AI study tools, a lockdown exam simulator, a knowledge bank and honest progress analytics — grounded in your own lectures, explained in plain English. - Branch research paths: Explore multiple angles on the same question, merge the best findings later - Read and cite: Open a notebook per name and cite the filing passage, the latest call and your prior notes, each claim beside its source. ### AI for lawyers — https://askfinz.com/for/lawyers Draft contracts and briefs in hours, not days — with every clause explained and every risk highlighted. - Study a case law: Open a notebook and cite the relevant filing passage, the latest call, and your prior notes. - Branch research paths: Explore multiple angles on the same question, merge the best findings later - Read the manual: Open the Docs workspace and follow the instructions to understand how to use askFinz effectively. ## Guides (70) ### Crawling, indexing, scraping: three words people use interchangeably — https://askfinz.com/guides/crawling-vs-indexing-vs-scraping Crawling, indexing and scraping get used as synonyms, but they're three different jobs. A plain-English breakdown of what each one actually does. "We crawled the site." "We scraped the data." "It's in the index now." In casual conversation these three phrases get swapped for each other constantly, and most of the time nobody notices, because the words are close enough that the sentence still makes sense. But they describe three different jobs, done by three different pieces of software, and mixing them up is the reason a lot of "why isn't my data showing up" confusion happens in the first place. Here is the plain version, in the order the work actually happens. ## Crawling: finding pages A crawler's job is discovery. It starts from… ### How to get your site indexed by AI search — https://askfinz.com/guides/get-my-site-indexed-by-ai Waiting to be discovered isn't the only option. How to get your site read, indexed and answerable by AI search — and how to keep control of it. If you run a website, "will AI search engines find me" used to be a question you couldn't do much about — you published, and you waited to be discovered whenever a crawler happened to reach you. That's still partly true. But it's worth understanding both halves of the picture: what happens automatically, and what you can actually do to speed it up or take more control over how your site is read. ## What happens without you doing anything If your site is public and reachable, it's the kind of thing a general web crawl is built to find in the ordinary course of working through links — no subm… ### Why a headless browser gets a 403 and a real one doesn't — https://askfinz.com/guides/headless-browsers-get-blocked Same machine, same cookies, wildly different result. Why headless Chromium gets refused where a real browser sails through — and what that means. Here's a scenario that trips up a lot of engineers the first time they hit it: you write a script using headless Chromium, point it at a page you can open fine in your normal browser, and it comes back with a 403. Same URL. Same cookies, if you copy them over. Sometimes even the same IP address. And yet one gets the page and the other gets refused outright. The instinct is to assume something about the request is wrong — a missing header, a bad cookie, the wrong referrer. Usually none of that is it. The real answer is blunter: the site fingerprinted the browser itself, decided it was automat… ### How a search index is actually built — https://askfinz.com/guides/how-a-search-index-is-built From a URL to an instant answer — a start-to-finish, plain-English walk through how a search index actually gets built and kept current. Typing a question into a search box and getting an answer back in a fraction of a second feels instantaneous, and that's exactly the point — none of the real work happens at query time. It happens beforehand, continuously, in a pipeline that turns the open web into something that can be searched instantly. Here's what that pipeline actually looks like, stage by stage. ## Stage one: discovery Before anything can be searched, it has to be found. A crawler starts from known pages and works outward by following links, building a map of what exists and how it connects. Coverage depends entirely… ### How AI search actually finds an answer — https://askfinz.com/guides/how-ai-search-finds-answers From your question to a sourced answer — how AI search bridges a live index and a language model to answer in words instead of ten blue links. Type a question into a traditional search engine and you get a page of links, ranked by relevance, and it's on you to open a few, read them, and work out the answer yourself. Type the same question into an AI search engine and you get an actual answer, in a sentence or two, with sources attached. The difference feels like magic, but the mechanism underneath it is well understood, and it's really just two established ideas working together in sequence: a search index, and a language model. ## It starts with an index, not a live web crawl The first thing worth clearing up is that AI search do… ### What web data actually costs: renting an index vs owning one — https://askfinz.com/guides/how-much-does-web-data-cost Every AI product needs web data from somewhere. The real cost difference isn't the sticker price — it's whether you rent an index or own one. Every AI product that answers a question about the current world needs web data from somewhere. Almost none of them read the web themselves — building and running that kind of infrastructure is a serious undertaking, so the overwhelming majority buy access to someone else's search results through an API and pass the cost (and the constraints) straight through to whatever they're building. Understanding that distinction is the actual key to understanding what "web data" costs, because renting and owning aren't two prices on the same thing — they're two different arrangements with very differen… ### Why most search can't read a PDF properly — https://askfinz.com/guides/indexing-pdfs-and-documents Slide decks, scans and reports hold some of the most useful information on the web — and most search engines link to them without reading a word. Try searching for a specific fact from a research report, a regulatory filing, or a company's investor deck, and you'll notice something odd: search engines will often find the document — they know it exists, they can show you the filename and the link — but they can't tell you what's actually on page 14. They found it. They didn't read it. This is one of the most underserved gaps in how search generally works, and it matters more than it looks like it should, because a huge share of genuinely useful information lives in exactly this kind of file rather than in an ordinary web page. ## Why… ### Real-time indexing vs scheduled crawling: what's the difference? — https://askfinz.com/guides/real-time-vs-scheduled-indexing Days to weeks behind, or searchable the moment it's read — the difference between scheduled crawling and real-time indexing, explained plainly. Ask most search systems how current their results are and you'll get a vague, reassuring answer. Ask more precisely — "if this page changed an hour ago, does your index know?" — and the honest answer, for a lot of systems, is no. Not because anyone is hiding anything, but because of how indexing has traditionally been built: on a schedule, not continuously. ## The scheduled model The traditional approach to keeping a search index current is to re-crawl on a fixed cycle — daily, weekly, sometimes longer for less important pages. A crawler works through its list, reads what's there, updates t… ### How robots.txt works for AI crawlers — https://askfinz.com/guides/robots-txt-and-ai-crawlers What robots.txt actually controls, how AI crawlers are supposed to read it, and how to check whether yours is doing what you think it is. A rising number of webmasters are asking a version of the same question: with AI crawlers now reading the web to build search indexes and train models, does my old `robots.txt` file still do what I think it does — and does it even cover the new crawlers showing up in my logs? The honest answer is: mostly yes, with some details worth getting right. ## What robots.txt actually is `robots.txt` is a plain text file, sitting at the root of your domain (`yoursite.com/robots.txt`), that states which parts of your site automated crawlers are and aren't welcome to read. It's not a technical lock — n… ### What is a vector index? Searching by meaning, not keywords — https://askfinz.com/guides/what-is-a-vector-index How a vector index lets a search system match ideas instead of exact words — a plain-English explanation of "searching by meaning." Search a keyword-based system for "how to stop feeling anxious before a talk" and it will look for pages containing those specific words. Miss the exact phrasing — the source you actually need says "manage nerves before public speaking" — and a keyword match can walk right past it, even though it's clearly the answer. A vector index exists to fix exactly this problem: it lets a system match on what something means, not on which words happen to appear. ## The old way: matching words Traditional search works by matching text. It builds something like an index at the back of a book — this word… ### What is a web crawler? — https://askfinz.com/guides/what-is-a-web-crawler What a web crawler actually does, how it decides where to go next, and how a well-behaved one is supposed to treat the sites it visits. Every search engine, every AI system that reads the live web, and every index of any real size starts with the same piece of software: a crawler. It's one of the oldest ideas in how the internet gets organised, and also one of the most misunderstood — mostly because "crawling" gets used loosely to mean almost any kind of automated visit to a website, when it actually describes something specific. ## What a crawler does A web crawler's job is to move through the web by following links, the same way a person would click from page to page, except automatically and at far greater scale. It star… ### Who owns the index behind your AI? — https://askfinz.com/guides/who-owns-the-index-behind-your-ai Most AI search features call a rented search index. Here's how to tell who owns theirs, why it sets their price, and what happens when it gets pulled. Ask an AI product where its web search actually comes from, and most of the time the honest answer is: not from them. The chatbot, the agent, the research tool — whatever sits on top — is usually a thin, well-designed layer over a search index someone else built. That's not a criticism; building and running a web-scale index is one of the harder infrastructure problems in software, and renting access to one is a completely reasonable way to ship a product. But it's also a fact worth knowing, because it quietly determines two things about whatever you're using: what it costs, and what happens… ### Why web scraping breaks (and keeps breaking) — https://askfinz.com/guides/why-web-scraping-breaks Selectors move, sites redesign, blocks appear overnight. Why scraped data pipelines fail so often — and what actually keeps working instead. If you've ever maintained a scraper, you know the pattern: it works fine for weeks, then one morning it returns nothing, or worse, returns garbage that looks fine until someone notices the numbers are wrong. Nothing about your code changed. The website did. This isn't bad luck or bad engineering. It's the structural problem with scraping as an approach — you're building against a target that has no obligation to stay still. ## The selector problem Most scrapers work by targeting a specific spot in a page's structure: "the price is inside this div, inside this class." That works exactly as… ### Agent-to-agent (A2A), explained — https://askfinz.com/guides/agent-to-agent-a2a-explained Agent-to-agent (A2A) lets AI agents hand off tasks to each other — enabling more complex, multi-step work than any single agent can handle on its own. A single AI assistant can do a lot. But some work is genuinely too large, too varied, or too parallel for one agent to handle end-to-end. Agent-to-agent (A2A) protocols define how AI agents can hand tasks to each other — breaking complex work into pieces, routing each piece to the agent best suited to handle it, and assembling the results. Understanding A2A helps clarify what is actually possible when AI systems are described as "agentic."