Project Brain

  1. › opening Project Brain
  2. › synthetic demo
  3. › loading workspace
Skip to content
Project Brainmemory for your AI agents

Workspace

DashboardProjectsMemoryQualityActivitySettingsSetup

Read-only example → install your own

Demo dataNumbers are examples until the API is connected.

Help

Workspace/Settings

Across all projects
Demo data · synthetic examples

Read-only preview. No repository or model is connected; the numbers and project knowledge are examples.

Install your own →

Loading this screen…

Settings

Alerts, scheduled jobs, search and indexing. Saving updates the API immediately. Background workers use the saved settings after they restart; a save does not restart workers.

GeneralAccessAgent toolsReranker

Loading measurements…

TelegramSchedulerActivity historyBackground checksSearch & context packsIndexing & workersModels

Telegram alerts

Send findings from background checks to your Telegram chat.

Last message

Delivered on 02 Oct 2026, 03:30 UTC

4 findings

Each problem is announced once, and again when it is fixed. Only a critical one that stays open is repeated this often.

Bot connection Configured

Create a bot with @BotFather, then add its token and the chat that should receive alerts.

Get a token from @BotFather. Leave blank to keep your saved token.
The chat or channel where the bot sends notifications.

Scheduled jobs

1 overdueBrain runs these jobs by itself. A job is overdue when it hasn’t succeeded for its interval plus 2 h.
Scheduled jobs
JobScheduleNext runLast successLast failureTries usedStatus
Nightly deep maintenancenightly_maintenance · deep queueDaily at 00:30 UTC07 Oct 2026, 00:30 UTC06 Oct 2026, 00:41 UTC—1/3ok
Nightly harness healthhealth_check · fast queueDaily at 02:00 UTC07 Oct 2026, 02:00 UTC06 Oct 2026, 02:00 UTC—1/3ok
Nightly LLM self-diagnosisself_diagnosis · fast queueDaily at 03:30 UTC07 Oct 2026, 03:30 UTC05 Oct 2026, 03:31 UTC
06 Oct 2026, 03:34 UTC

LLM provider timed out after 120s (attempt 3/3)

3/3last run failed
Weekly smoke benchmarkbenchmark · maintenance queueMondays at 06:00 UTC12 Oct 2026, 06:00 UTC28 Sept 2026, 06:04 UTC—1/3overdue

Activity history

  • Off by default. Capture up to 500 characters of future queries. Queries can contain source or secrets; disabling does not erase previously captured text.

Background checks

  • Brain checks its own health on a schedule and reports findings.

  • Off = deterministic probes only, no model calls.

  • Scheduled scan for risks and drift in the indexed code.

  • Nightly re-scoring and quality checks of the index.

Search & context packs

  • Better ordering of search results, slower and costs model calls.

  • Faster briefings without the model’s plan step.

1–50
Upper bound on files picked for one briefing.

Indexing & workers

1–32
Higher is faster on big machines.
1–16
Keep low on rate-limited providers.
1–256
Texts per embeddings request.
60–21600
A background job is stopped after this long.

Models

Change
Language model
openai gpt-4.1-mini
Embeddings
openai text-embedding-3-small · 1536 dims

Connect the Brain API to run actions — this is demo data.