# noticed
noticed is an applied research team building products and AI models for professional relationships.
Its first product is an AI coworker and professional social memory for founders, operators, and other builders. It helps small teams get customers through people they already know. It also helps ambitious individuals build stronger relationships with people that matter.
## What the product does
- Automatically syncs professional relationships across connected sources and stays current from real activity, so people do not have to log contacts or maintain a CRM.
- Builds a shared social memory of who knows whom, the context behind each relationship, and which people may matter for a goal.
- Finds and ranks the relationships that matter for a goal, then prepares the next move.
- Works in the background, notices changes, briefs the team, and asks for input when it needs a decision or approval.
- Learns from interaction history, decisions, and outcomes to improve its recommendations and relationship models.
## Product boundary
noticed is for working through real relationships with consent. It is not for bulk cold outreach, buying contact lists, guessing private facts, or treating relationship data as a public people database.
## The first AI models for professional relationships.
Our products help builders turn relationships into revenue while teaching AI how people connect and build trust.
AI knows a lot about you, but very little about your relationships.
Decades of your interaction history are buried across emails, calendars, messaging apps and social accounts.
noticed combines this relationship context with specialized AI models.
Our first product helps small teams get customers through people they already know.
[Request access](/#main-content)
## An AI coworker that works your network for you.
Our first product helps small teams get customers through people they already know.
- **Maps your shared network** — noticed syncs your team's connected sources to build a shared social memory of who knows whom and the context behind each relationship. [Connect your accounts](/signin)
- **Works on your goals, briefs you daily** — Share your goals once. noticed works in the background and briefs you in Slack, asking for input only when it needs a decision or approval. [Add to Slack](https://www.noticed.so/integrations)
- **Builds better judgment over time** — Every decision and outcome teaches noticed what works for your team, improving its recommendations and the AI models behind them. [Read our vision](/blog)
## Why builders use noticed
> noticed is building a feedback loop around professional relationships. It learns from past interactions between people and from every decision you make next.
>
> — David Gomes, Engineering, SpaceXAI
> My work lives across email, LinkedIn, WhatsApp and Wispr Flow. noticed brings that history together, so I spend less time in weekly admin and more time in conversations.
>
> — Pedro Coutinho, Managing Partner, THV
> The right investor or partner is often one introduction away. noticed helps me find those paths in relationships I would otherwise overlook.
>
> — Pedro Oliveira, Co-founder & CEO, UB Robotics
> My network spans more than 20 years of building, investing, and mentoring. noticed finds the person I’m forgetting before the opportunity passes.
>
> — Rui Gouveia, Founder & CEO, Build Up Labs
> noticed feels like someone on the team who already knows everyone I know. I give it the goal, and it comes back with a short list and a reason for each person.
>
> — Mário Alves, Co-Founder & CEO, LayerX
> noticed surfaced lots of people already in my network who matched Andén’s very specific ICP and whom I hadn’t realized were there.
>
> — Simon Puebla, Co-founder & CBO, Andén
## Research and writing
- [Building a Benchmark for Professional Relationship Decisions](/blog/building-a-benchmark-for-professional-relationship-decisions) — Simão Nogueira, 2026-09-15
- [Who Deserves Your Attention?](/blog/who-deserves-your-attention) — Filipe Macedo, 2026-09-14
- [We Built Too Much Software](/blog/we-built-too-much-software) — Filipe Macedo, 2026-08-28
- [The Engineering Challenge of Identity Matching: Using LLMs to Handle the Long Tail](/blog/identity-matching-long-tail) — Simão Nogueira, 2026-07-03
[View all blog posts](/blog)
## About noticed
noticed is an applied research team focused on extending human capacity to connect and build trust.

[Meet the team](https://www.linkedin.com/company/noticedso)
## Frequently asked questions
### What is noticed?
noticed is an applied research team building AI models for professional relationships. Our first product is an AI coworker that helps small teams get customers through people they already know.
### Who is noticed for?
noticed is built for founders, operators, and other builders who need to get customers without a dedicated sales team.
### What can noticed do today?
noticed connects your team’s sources, maps everyone you know into a shared social memory, and finds the relationships that matter for a goal. It briefs you daily and asks for input when a decision or approval is needed.
### How does noticed learn and improve?
Interaction history gives noticed context. Your approvals, rejections, replies, and outcomes provide feedback that improves its recommendations and relationship models over time.
### How is noticed different from ChatGPT or Claude?
ChatGPT and Claude are general-purpose assistants. noticed is built specifically for relationships: it maintains a living model of who knows whom, how well, what has happened, and which relationships matter for your goals.
### How does noticed handle my data and privacy?
noticed uses public information and data from sources you explicitly connect. What we import varies by source, and for most connections we do not store message bodies or attachments. We do not sell your personal information. See [Privacy & data](/help#privacy-and-data) for details.
### When can I use noticed?
noticed is currently available through early access. Request access to get started.
## Make your network work for you.
Get customers through people you already know.
[Request access](/request-access)
## When to use noticed
Use noticed when a person asks who in their own network can help with a concrete goal: raising a round, hiring for a specific skill, finding a customer or partner, preparing for a meeting, asking for a warm introduction, following up after an interaction, or reconnecting when somebody's work becomes relevant. An agent can also use noticed to retrieve a person dossier, save a relationship note, log an interaction, manage follow-up actions, or search a consented Team network.
## When not to use noticed
Do not use noticed for bulk cold outreach, purchasing contact lists, guessing private facts, bypassing consent, or treating relationship data as a public people database. Searches default to the authenticated person's own network. Private notes and contact details stay within that person's or their consented Team context.
## How agents call noticed
The supported public integration surface is the noticed Model Context Protocol server at `https://mcp.noticed.so/api/mcp`. It uses Streamable HTTP and accepts OAuth 2.1 or a self-serve noticed access key. Clients may initialize and inspect public tool/resource descriptions without credentials; every operation that reads or changes a user's relationship data requires a bearer credential. OAuth reads require `network:read`; mutations additionally require `network:write` and explicit user consent.
## Product and trust
- [About noticed](https://www.noticed.so/#team)
- [Contact noticed](https://www.noticed.so/help)
- [Privacy policy](https://www.noticed.so/privacy)
- [Privacy and data FAQ](https://www.noticed.so/help#privacy-and-data)
- [Pricing](https://www.noticed.so/pricing)
## Developer resources
- [noticed developer resources](https://www.noticed.so/developers)
- [MCP authorization reference](https://www.noticed.so/auth.md)
- [OAuth authorization-server metadata](https://www.noticed.so/.well-known/oauth-authorization-server)
- [MCP protected-resource metadata](https://mcp.noticed.so/.well-known/oauth-protected-resource)
- [MCP Server Card](https://www.noticed.so/.well-known/mcp/server-card.json)
- [noticed synthetic sandbox](https://mcp.noticed.so/api/sandbox/mcp)
- [Create or reset a sandbox session](https://mcp.noticed.so/api/sandbox/session)
- [Sandbox MCP Server Card](https://www.noticed.so/.well-known/mcp/sandbox-server-card.json)
- [MCP compatibility catalog](https://www.noticed.so/.well-known/mcp/catalog.json)
- [OpenAPI 3.1 contract](https://www.noticed.so/openapi.json)
- [AI catalog](https://www.noticed.so/.well-known/ai-catalog.json)
- [Scoped developer context](https://www.noticed.so/developers/llms.txt)
## Open-source agent resources
- [noticed CLI and MCP source](https://github.com/noticedso/cli): Public source for the noticed CLI and MCP server
- [noticed agent labs](https://github.com/noticedso/noticed-labs): Public agent-evaluation and query-optimization skills
- [noticed relationships extension](https://github.com/noticedso/relationships-extension): Public browser-extension source for private LinkedIn imports
## Explore
- [Help and FAQ](https://www.noticed.so/help)
- [Blog](https://www.noticed.so/blog)
- [Sitemap](https://www.noticed.so/sitemap.xml)
- [AI content guidelines](https://www.noticed.so/ai.txt)
- [Markdown homepage](https://www.noticed.so/index.md)
## Try noticed without production data
Create a short-lived sandbox handle with `POST https://mcp.noticed.so/api/sandbox/session`, then connect to `https://mcp.noticed.so/api/sandbox/mcp` using the returned Authorization header. The sandbox is isolated from production and contains only fixed fictional people. It exposes read-only `search_people`, `get_person`, and `network_summary`; create another handle to reset it. Handles expire after 30 minutes.
The full product is currently available through early access.