Bcontext

Building the brain your company always needed.

Start now

Free while we’re in beta. We set every organization up by hand — usually within a day.

The problem

Your agents start every session from zero.

Every time you open Claude Code, Cursor or a chat, the agent knows nothing about your company: what you decided, what is in flight, why. You explain it again, and your team pays for it in time and tokens.

When the session ends, that context is gone. Two people and two agents make four versions of the truth, and the right one lives in a Slack thread or in someone’s head.

monday · claude code · Ibon
› what did we decide about payments?
I don't have any context about your company.
Could you walk me through the project?
tuesday · cursor · Ibon
› what did we decide about payments?
I don't have any context about your company.
Could you walk me through the project?
wednesday · claude code · María
› what did we decide about payments?
I don't have any context about your company.
Could you walk me through the project?

What Bcontext is

A shared memory your team and your agents read and write.

Bcontext is your company’s working memory: decisions, tasks, docs and goals, linked to each other. Your team uses it from the web. Your agents connect once over MCP and, from then on, read what they need and write down what they do, with the permissions you give them.

Everyone else builds memory the agent reads. We built the one the agent writes.

AAcme
Dashboard
Nexo
Knowledge
Brain
Tasks
Roadmap
Browse
Agents & Tools
Agents
Tools
Brain41 nodes · 63 links
decisiontaskdocgoal

Day to day

brief when you start, Result when you finish.

  1. You open a session

    The agent calls brief and gets what is open, what was decided and what changed since its last visit. You explain nothing.

  2. You work as usual

    It reads the decisions and docs it needs, creates tasks, and leaves a record of what it did.

  3. You close

    It writes a Result. The next session — yours, a teammate’s or another agent’s — starts knowing it. And your team sees it happen.

› brief
open        1  Move payments to Stripe Link       high
            2  Finish the onboarding copy           med
decided     Passkeys replace magic links
since yday  María closed «Passkey auth»            Result
            cursor-1 opened «500 on /api/export»
› starting with 1
AAcme
Dashboard
Nexo
Knowledge
Brain
Tasks
Roadmap
Browse
Agents & Tools
Agents
Tools
decision

Stripe Link instead of our own card form

Context

38 % of carts were abandoned at the card step. We tested Link on 20 % of traffic for two weeks.

Decision

Stripe Link for every customer from October 1. The in-house form is retired, not kept “just in case”.

Result

Deployed to production. Abandonment at the payment step: 38 % → 21 % in the first week.

written by claude-code · 2 min ago

Who it's for

Teams that already work with agents.

  • Teams of 1 to 10 people.
  • Claude Code, Cursor or similar every day.Bcontext shows on the agent’s second session. No agents, no second session.
  • Several tools and more decisions a week than fit in Slack.
  • Founders who want a say in how the tool they use gets built.

Not for you yet if your team doesn’t use agents daily, or if what you want is a wiki or a task manager. You already have those.

AgentPermissionsLast seen
ccclaude-code · Ibonread · write · tasks2 min ago
ccclaude-code · Maríaread · write · tasks40 min ago
cucursor-1read · writetoday, 09:15
cici-reviewerreadyesterday
nxnexoread · ingest5 min ago

Founding Users Program

Thirty days to find out if it’s yours.

  • A 30-day pilot with the whole product, at no cost.
  • Onboarding with us on day one: we connect your agents, load your repo and docs, and run the first session with you.
  • A direct line during the pilot, and founding-user conditions if you stay.
  • In return: real use, honest feedback, and thirty minutes a month.

After thirty days you decide. If it doesn’t earn its place, you stop — and you will have taught us something.

Start now

Free while we’re in beta. We set every organization up by hand — usually within a day.

Day 1

Onboarding with us

Plugin installed, agents connected, repo and docs loaded, a first real session with decisions written down.

Day 2

The second session

The agent opens with brief and already knows about yesterday. This is where it shows, or doesn't.

Weeks 1–4

Real use

The whole team, every agent. We watch usage and fix whatever slows you down.

Day 30

Your call

Keep going as a founding user, or stop.