ai & agentsproduct of a company I co-founded
Footprint: company reports from agent crews.
A platform for composing agent pipelines from the interface: each agent with its own model and plugins, in parallel or in sequence. At the end of the chain comes a company report, built, reviewed and translated.
The project in numbers
Composable pipelines
agents in parallel or in sequence
One model per task
chosen for what each does best
Plugins and MCP (Model Context Protocol)
tools added when they are needed
the system, in cross-sectionfrom brief to report
the crew, in planwho works, and in what order
A pipeline is composed, not programmed.
The agents sit in parallel where the work is independent and in sequence where the previous step’s result is needed. Each brings the model suited to its task and the tools it needs.
brief
Client request
in natural language, from the interface
in parallel · one model per task
Research
fast model, many sources
Analysis
reasoning model
Data
internal plugins and MCP
in sequence
Assembler
merges the contributions into one text
Review
critical analysis and translation
Report delivered
ready to read
what changes
Pipelines are built in the interface or via API: adding an agent does not mean rewriting the system.
Whoever uses Footprint composes their own flows, without depending on me for every change.