Your agent does the research. You keep the receipts.
Bring your own Claude or GPT key and point it at 19,000+ companies — theses, signals, and reports, drafted by agents you direct. Every figure the tools return traces to the filing it came from, and anything irreversible waits for your approval.
Drafting the brief into the document beside you.
reading the FY2024 10-K…
NVDA — Q1 research brief
$130.50B
└─ 10-K · fact_id 4c1f…9ab2
Bring your own model. We bring the evidence.
Most research tools lock you into their model, their UI, and their pricing. The Workspace flips that: your Claude or GPT key drives the chat, your prompts run our playbooks, and every answer is grounded in live SEC filings — not the model's memory.
Your key. Your model. Your bill.
Paste an Anthropic, OpenAI, or OpenRouter key — sealed with AES-GCM, plaintext never stored, cookie-only mode available. Switch providers anytime. We never train on what you ask.
Wired to 121 SEC tools.
Every chat session has live access to fundamentals, ratios, peer comps, smart-money flow, forensic audits, DCF compute — the same MCP that Claude Desktop and Cursor users connect to directly.
From question to artifact in one tab.
Chat → save thesis → schedule signal → publish report. No copy-pasting between five apps. Everything you create is yours and shareable on your terms.
One workspace. Every surface.
Each feature is live today. They share one Bearer token, one MCP, and one identity — so what you build in chat automatically lands in claims, theses, reports, portfolios, watchlists, signals, and scheduled automations.
BYO-LLM Chat
Bring your own Anthropic, OpenAI, or OpenRouter key — sealed with AES-GCM, plaintext never stored, and the stored copy for scheduled agents removable in one click. Or skip the key and use the managed Valuein AI lane. Either way it's wired live to all 121 MCP tools, so every answer is cited.
Findings
The atomic, evidence-backed unit: one falsifiable statement, cited to a filing, auto-graded against the data (confirmed / refuted). Compose findings into theses — improve a finding and every thesis and report built on it improves.
Investment Theses
Save directional bets (bull / neutral / bear) with conviction (1–5), horizon, and notes — composed from your claims and auto-graded against forward returns into a verifiable track record.
Reports
Author research in a Tiptap editor with autosave + version history. Publish to a shareable /r/[slug] URL with private / unlisted / public visibility. Citations attached to every claim.
Show the full workspace
Portfolios
A private book of holdings under a written mandate. Compose your theses, claims, and watchlists by ticker — then seed a scheduled agent straight from a portfolio's ticker set.
Watchlists
Group up to 500 tickers per list, with optional screening criteria. One click to wire a watchlist into a filing signal or kick off a daily brief.
Get Notified → Inbox or Agent
Five conditions (filing events, ratio thresholds, price moves, fundamental changes, watchlist changes) delivered to your inbox bell, email, or an HMAC webhook — or fire an agent the moment the condition trips. Test-fire before you rely on them.
Agents
Build a deterministic Recipe or an agentic Team, run it on demand, or put it on a Schedule that fires on cadence. Every run lands in one Activity timeline you can review, each keeping the research file that shows how it got there.
Scheduled Tasks
Your agent defers its own work: "re-check AAPL margin compression in 30 days" becomes a durable task that wakes on time and re-runs the research for real — not a reminder, a run.
Signals
Trigger → action, generalized. A filing landing, a price move, or a signal firing can kick off an agent, send a signal, or post to your inbox. Dry-run any signal before you enable it.
Staged Approvals
Human-on-the-loop by construction: read-only tools run free, but mutating and high-impact actions stage for your explicit approval — every decision recorded in an audit entry.
39 Playbooks + 121 Tools
One-click flagship workflows — equity research brief, screen & shortlist, capital allocation, smart-money screen, forensic audit — plus the full MCP tool catalog, searchable and droppable straight into chat.
Morning Brief
Wake up to a daily brief on your watchlist — what filed overnight, what to read, what to watch. Grounded in live tool calls on paid tiers, generated on your own model, and readable from any MCP client too.
Profiles & Publishing
Claim a vanity @handle, then publish theses and reports to SEO/AEO-indexed profile + /r/[slug] pages. A verifiable claim + thesis track record builds your reputation — private by default.
How it works
You bring the agent. Valuein's MCP brings point-in-time SEC data as tools, playbooks, and workflows — your model, your key, our data.
Valuein MCP Server
mcp.valuein.biz
point-in-time · survivorship-free
The 15-minute setup
Do them in order. Steps 1–3 are the “aha”; 4–7 turn it into a desk that keeps working while you sleep.
- 01
Sign in & pick a handle
Sign in with Google, LinkedIn, or GitHub. On first sign-in you pick a public handle — valuein.biz/@yourhandle. One-time, required before you publish.
- 02
Add your LLM key (or use Valuein AI)
Paste an Anthropic, OpenAI, or OpenRouter key in chat — sealed with AES-GCM in a 24-hour httpOnly cookie, plus an encrypted vault so your scheduled agents can run while you're away. Switch the vault off in chat settings to keep the key in the cookie only. Prefer no key? Use the managed Valuein AI lane.
- 03
Run your first query
Type: "Give me a quick research brief on NVDA." Cited, point-in-time facts come back in seconds.
- 04
Create a watchlist
Drop in 5–20 tickers you actually care about. This becomes the spine of your routine.
- 05
Set one signal
Watchlist change → Inbox. Now any new filing from those names lands in your bell automatically.
- 06
Record one thesis
Your highest-conviction name, composed from a finding or two. This seeds your track record and starts auto-grading.
- 07
Connect your desktop AI
In Account settings → API, copy the Claude Desktop / Cursor / Codex config with your token pre-filled, then hit Test connection.
Pick your playbook
Choose the role closest to you: a one-time setup, then a daily, weekly, and monthly routine. Every step is something you type into chat.
Equity Analyst
You cover names, prep earnings, and ship memos. The workspace turns coverage into a system that self-updates.
- Build a "Coverage" watchlist of every name you follow.
- Add a filing-event signal on those tickers (8-K / 10-Q / 10-K) → Inbox, plus email for your top 5.
- Open a thesis on each covered name so your view is on record and self-grades.
- Clear the Inbox — new filings on your coverage.
- Ask chat to summarize any 8-K and whether it changes the thesis.
- Run equity_research_brief (full depth) on the 1–2 names in focus → save as a Report.
- Before earnings, pull the last 8 quarters and read the trend and management credibility.
- Publish one report publicly — your weekly credibility deposit.
- Run a multi-year operating + moat read on a core holding.
- Review capital allocation — is management creating value with retained earnings?
- Re-score theses past their horizon and refresh your views.
Start anything by typing it in chat: "run equity_research_brief on AAPL, full depth".
Upgrade unlocks — Pro unlocks the full 19,000+ universe, your public analyst profile, and one-click DOCX/XLSX deliverables for clients and IC. Institutional adds the smart-money layer, intraday filing webhooks, and full history back to 1993.
Use it alongside MCP, not instead of it.
Same Bearer token, same 121 tools, same point-in-time data. Pick the surface that fits the moment.
MCP — your agent
For Claude Desktop, Cursor, Codex CLI, custom agents. You own the UX; we power the data.
- Streamable HTTP, OAuth 2.1 + PKCE
- Zero workspace overhead — pure tool access
- Best for: power users, dev environments, agent platforms
Workspace — our chrome
For analysts who want chat + theses + reports in the browser, without configuring a thing.
- Chat + 39 playbooks + theses + watchlists + signals + reports
- Public profile + reputation scoring
- Best for: analysts, PMs, and individual investors
Frequently asked
Is the workspace production-ready?
Yes. Each card above shows its own live status — nothing here is a mockup or a coming-soon placeholder — and everything runs on the same infrastructure as our API. We ship improvements continuously. Bug reports and requests go to support@valuein.biz.
Do agents run on their own, or do I stay in control?
You stay on the loop. Read-only research runs freely, but mutating and high-impact actions stage for your explicit approval before they execute — approve or reject from the Approvals page, and every decision lands in an audit entry. Managed runs use a model allow-list with destructive tools stripped, and the agent is instructed to take every figure from a deterministic tool — a number without a fact_id didn't come from Valuein. There is no fully-autonomous mode.
Do you store my Anthropic or OpenAI API key?
Never in plaintext. For interactive chat your key lives in an AES-GCM-encrypted httpOnly cookie that expires in 24 hours. So your scheduled agents can run while you're away, the encrypted ciphertext is also kept in a key vault by default. Switch the vault off in chat settings and the key lives only in that cookie (re-enter once a day, nothing in any database). The vault can only be decrypted with a secret held outside the database, removal is one click and immediate, and your key is never logged and never used to train anything.
How is this different from connecting Valuein MCP to Claude Desktop?
Same MCP server. Same auth. Same 121 tools. The workspace is the alternative when you don't want to leave the browser or wire up Claude Desktop — it ships chat, claims, theses, reports, portfolios, signals, and scheduled automations on top. Use MCP from your own agent or use ours. Or both.
Are my theses and reports public by default?
No. Default is private. Each thesis and report has a visibility flag: private (only you), unlisted (anyone with the link), public (indexed by search engines and AI agents at /your-handle/...).
Which tier do I need?
The free Benchmark tier (just register, no card) unlocks the workspace scoped to S&P 500 constituents — create theses, signals, reports, and run the full BYO-LLM chat. Pro ($49/mo) and Institutional ($499/mo) widen scope to the full 19,000+ universe — 15-year rolling history on Pro, full archive back to 1993 on Institutional — plus public profile + reputation scoring, and (Institutional) the smart-money tables. Paid plans are single-seat; teams and firm-wide use are licensed by contract (Enterprise), and redistribution needs a separate Commercial licence. The Sample tier (no signup) is read-only — it can browse public profiles and shared reports but not create them.
Ship your first thesis tonight.
Sign up free, claim your handle, paste your Anthropic or OpenAI key. Ten minutes to first claim. Private by default — public, unlisted, or shared on your terms.
Or see pricing.