Every team goes through the same stages. The first three are the infrastructure shift happening now. The last one is what that foundation eventually unlocks.
From one agent
to an engineering floor that never clocks out.
You started with Claude Code. Then worktrees. Now your agents and engineers share cloud workspaces backed by real environments, durable sessions, and clean handoff. Palmer is the workspace layer that makes serious engineering agents possible. Volition comes later, once the environment is in place.
AGENT-01
REVIEWING
customer-tickets - scanning latest batch
AGENT-01
FIXING
ticket-4821 - auth redirect loop on mobile
AGENT-01
COMPLETED
8/12 tickets resolved - 4 PRs merged
AGENT-02
SYNCING
granola - pulling meeting action items
AGENT-02
BUILDING
onboarding-v2 - implementing new user flow
AGENT-02
SHIPPED
PR #312 merged - onboarding redesign live
AGENT-03
ANALYZING
eval-suite - running accuracy benchmarks
AGENT-03
IMPROVING
prompt-tuning - accuracy 84% → 88%
AGENT-03
VERIFIED
accuracy 91% - 7 point improvement
▊
Scroll
01 The Ladder
The progression from
The progression from
one agent to full autonomy.
One Agent.
Claude Code in your terminal. One task at a time. You prompt, wait, review. Powerful, but sequential. Your machine, your attention, your bottleneck.
"What if you could run dozens at once, each with its own environment?"
Many Agents.
Cloud workspaces that own the work: sessions, artifacts, approvals, and handoff. Each workspace can spawn one or more real environments with full database, CI, and test access. Run dozens in parallel and see exactly which sessions are blocked and why.
"What if they could keep working after you leave the office?"
Overnight.
Agents keep running while you're not looking. They test, verify, and do QA in real environments, and your team can reopen the same workspace in the morning to inspect, take over, or hand it to someone else. Engineering output decoupled from office hours.
"What if they could figure out what to work on themselves?"
Volition.
Once environments and workspaces are in place, you can ask for more: agents that decide what to work on next. They pull from customer tickets, meeting notes, and eval results. They prioritise, investigate, build, and ship. Monday morning, everything is done.
"This is what comes after the workspace layer is in place."
02 Real Environments
Same stack as
Same stack as
your engineers.
Every task runs inside a real cloud environment. Not a sandbox, not a container with half your dependencies missing. The same stack your team develops against, spun up on demand with the repos, tools, network access, and scoped credentials the work actually needs.
This is why the output is high-quality. Agents run the full test suite, check their builds, validate in staging, and can even use computer-use models to do product QA before opening a PR. Your team can log into that same workspace to inspect the result manually. No black box.
Environment is where code executes. Workspace is where work accumulates.
03 Perfect Context
We embed with your org.
We embed with your org.
Your workspaces inherit that context.
Palmer doesn't guess at your codebase from a README. Our engineers embed with your team, learning your architecture, your conventions, your priorities, and how work actually moves. Then that context gets bound into your workspace definitions, so every session and environment run starts from the same map.
Have something custom? We'll integrate it.
Granola Meeting Notes
Q3 planning sync. Action items: redesign onboarding, add SSO support, improve mobile performance
Onboarding redesign implemented. PR #312 shipped. SSO scaffolding ready for review. Mobile perf audit in progress.
Zendesk Customer Tickets
12 new tickets overnight: auth issues, slow dashboards, mobile redirect loop, missing exports
8/12 resolved. 4 PRs merged. Remaining 4 flagged. Need product decisions before proceeding.
Langfuse Eval Results
Accuracy dropped to 84% after prompt changes. Investigate root cause and improve.
Root cause identified. Prompt tuning complete. Accuracy: 91%. Seven point improvement, verified across full eval suite.
04 Pricing
Start with a team.
Start with a team.
Scale to a floor.
Self-Serve
$100
per month
Bring your own Codex or Claude subscription. Get real environments and durable workspaces on your own terms.
- Bring your own Codex / Claude
- Real cloud environments
- Durable workspaces
- Bring your own agents
- Community support
Startup
$2,000
per month
For teams under 10 engineers. We set up the workspace layer: environments where code runs, and workspaces where sessions, artifacts, and handoff live.
- Engineers embed with your team
- Setup and onboarding included
- Real cloud environments
- Shared workspaces for humans and agents
- Monthly consulting call
Enterprise
Custom
contact us
Everything in Startup, plus deeper integrations, dedicated support, and custom workspace policies and environment templates for larger teams.
- Dedicated support engineer
- Custom integrations
- SLAs and compliance
- Priority access to new features
General intelligence is the ability to solve any problem.
General volition is the ability to decide which problems are worth solving.
Built by engineers from
Citadel
Bloomberg
Former Forward Deployed Engineers and quantitative engineers.
London.
We're hiring engineers who want to work on what might be the most interesting open question in AI. People who are equally comfortable reading research papers and shipping production code.
Careers:
jobs@generalvolition.com
Early access:
hello@generalvolition.com