Coming Soon

AI Engineering
Conductor

Vector coordinates specialist AI agents, multiple models, and your developer tools so AI can finish real engineering tasks — not just suggest lines of code.

Vector — conductor
You Fix the race condition in the auth middleware and add a regression test.
Planner Analysing auth/middleware.ts and recent incident history...
──── Plan ready · 3 steps · approve to start
1. Isolate token refresh lock in middleware
2. Patch concurrent refresh guard
3. Add concurrency regression test
✓ Code Agent Patch applied · diff ready for review
✓ Test Agent 3 tests added · all passing
✓ Review Agent No issues found · ready to commit
>

Built for how engineers actually work

A focused environment — not another chat window. Sessions persist, agents run in the inspector, and every action is visible before it lands.

Conductor — conductor · main
C
Conductor conductor · main
Auth middleware fix 14:23
Add rate limiting 11:05
Refactor token store 09:41
conductor / Auth middleware fix ● Local
You

Fix the race condition in the token refresh middleware and add a regression test.

Conductor

Analysing auth/middleware.ts — identified concurrent refresh guard issue. Plan ready.

Plan · 3 steps
✓ Isolate token refresh lock
✓ Patch concurrent guard
⟳ Add regression test
Ask Conductor to build, debug, review…
Activity Changes Context
Status⬤ Running
ProviderAnthropic
Modelclaude-sonnet
✓ Isolate refresh lock code
✓ Patch guard code
⟳ Regression test code

Agents that finish tasks

Not just autocomplete. Vector understands your project, routes work to the right model, and keeps you in control.

🗺️
Live Project Graph

Persistent understanding of your codebase — files, relationships, dependencies, tests, docs, and recent history. Always current.

🔀
Model Router

Routes each task to the best model for the job. Fast model for search, reasoning model for planning, local model for private files.

🤝
Approval Gates

High-impact actions require your sign-off. File edits, terminal commands, and deployments never happen without your approval.

🔍
Code Conductor

Plan, implement, review, and test changes end-to-end. See the diff before anything touches your repo.

🚨
Incident Conductor

Paste a stack trace or alert. Vector searches commits, deploys, and code to find root cause before you arrive.

🔒
Privacy First

Local-only mode. Bring your own API keys. No code sent anywhere without your explicit consent. UK/EU data residency option.

The right agent for every task

Vector coordinates a team of specialist agents, each with a defined role and scope. Coming progressively in each release.

Planner
Code
Review
Test
Security
Incident
Migration
Docs
Onboarding
Standards
Workflow
Model Router

Works with any model

Route work to the best provider for each task. Never locked in to one vendor.

Claude GPT-4o Gemini Mistral Ollama (local) LM Studio Self-hosted + more

Built for privacy-conscious developers

Vector is designed as a UK/EU-friendly AI development platform. Your code never leaves your machine unless you choose it.

Structured consent-based data collection means you decide exactly what improves the system — and you can opt out entirely.

  • Local-only indexing mode
  • Bring your own API keys
  • No code sent without explicit consent
  • Secret scanning before any upload
  • Project-level and org-level privacy settings
  • UK/EU data residency option (planned)
  • Self-hosted enterprise deployment (planned)
  • Full audit logs

Building in the open

Phased delivery. Each release proves a distinct capability before moving to the next.

Phase 1 Desktop foundation — Tauri shell, SQLite, persisted sessions, project and provider views Done
Phase 2 Developer MVP — Live Project Graph, LLM routing, Planner Agent, Code Agent, multi-provider support In progress
Phase 3 Team MVP — GitHub integration, PR review, ticket-to-PR workflow, shared project memory Upcoming
Phase 4 Incident & Security — Incident Conductor, log investigation, CVE handling, rollback recommendations Upcoming
Phase 5 Automation — Workflow builder, scheduled agents, multi-agent pipelines, Migration Conductor Upcoming
Phase 6 Sovereign AI — UK/EU data residency, opt-in evaluation datasets, enterprise self-hosted deployment Upcoming

Simple, transparent pricing

Bring your own API keys — pay only for Vector, not per token. Cancel any time.

Hobby
Free

Try Vector with local models at no cost.

  • 1 workspace
  • Local models only (Ollama)
  • Planner Agent
  • Persisted sessions
  • Community support
Join waitlist
Solo
£12/mo

For individual engineers shipping with AI.

  • Persisted sessions & conversations
  • 3 AI providers (Anthropic, OpenAI, Ollama)
  • Model router — Simple / Standard / Complex
  • Planner Agent + Code Agent
  • 3 active workspaces
  • Local-only mode (BYOK)
Join waitlist

7-day free trial — no card required

Studio
£89/mo

For teams and engineers who need custom tool integration.

  • Everything in Pro
  • MCP integration — connect any tool via Model Context Protocol
  • Team seats (up to 5)
  • Shared session history
  • UK / EU data residency
  • Dedicated Slack support
  • Early access to new agents
Join waitlist

All plans require your own LLM API keys — we never mark up model costs. Prices in GBP ex. VAT.

Early access coming soon

Vector is in private development. Join the list to be notified when early access opens.

No spam. Unsubscribe any time.