private beta · for deal teams
Edit the plan before it runs.
Hand it a complex task. It proposes the agents, the models, the dependencies. You change anything that doesn't fit your house — then run it.
You set the budget cap · workspace-isolated · every run audit-trailed
↓ try it — this is the actual plan editor
proposed plan · 1 wave · 4 agents
editable
wave 1 · parallel
wave 2 · after wave 1
estimated total ~$0.78 of $1.00 cap
it gets better every time you use it
Most AI tools start from zero. This one remembers how you work.
Run a task you'll repeat, and it offers to save it as a skill. From then on, similar tasks start from your template — your agents, your models, your conventions. Your edits and feedback feed back in, so each run is sharper than the last.
first run
$0.65 · 3 edits before approval
Generic plan. You renamed two agents, swapped a model, tightened the budget. Output is fine but in the system's voice.
tenth run, same task type
$0.48 · 0 edits
Plan loads from your skill. Cheaper because the planner skips ground it's already covered. Output reads in your house style.
how it works
Stage 1
You write the prompt.
Screen Meridian Cold Chain for our IC. Read the CIM, run comps, flag risks, draft a one-pager.
Stage 2
It proposes a plan — you edited it above.
Rename agents, swap models, add steps, tighten budget. The plan that runs is the one you signed off on.
Stage 3
Independent agents run in parallel — dependent ones wait their turn.
researcher_financials
running…→ read_pdf("meridian_cim.pdf", pages="12-28")
researcher_comps
done · $0.092 of 4 done · $0.18 of $1.00 cap
Stage 4
You get the finished deliverable.
Meridian Cold Chain — IC screening memo
Thesis. Refrigerated last-mile logistics platform serving regional grocery and pharma. Three-year revenue CAGR of 22.7% to $84M, EBITDA margin expanding from 11.4% to 18.0% on route density.
Key risks. Customer concentration — top three accounts are 31% of revenue; one is on a month-to-month MSA. Ask multiple of 8.3x 2025E EBITDA prices in continued margin expansion that depends on the new Memphis hub clearing utilization above 72%.
+ comp set (4 firms), 8 questions for management, full source citations.
Built by Matteo Craviotto — MS Financial Engineering @ USC email me.