The Real Jaseci Products
Jac-GPT - Interesting agent 2 -> (Open source, enterprise edition (Open source, dashboard configure integration, onprem tooling)-- Retool JacGPT -> AnyGPT
Dashboard/configurator/integration-- Desktop App, white-label customizable-- Auto specialize
-- JacGPT -> AnyGPT and Open source launch
-- Laucn MVP Enterprise edtion (autospecialize not required)
-- Autopecialize is priority in AnyGPT roadmap, we build accordingly
Autospecialize IP
Special Push button auto-specialize:
Step 1: use opus to get many question from docs
Step 2: Use opus to run agent on questions to generate dataset
Step 3: Use dataset to train SLM to be as good as opus on docs
Jac coder semi auto specialize
Step 1: Run product with opus for a while to get data
Step 2: Use dataset to train SLM to be as good as opus on docs
Kimi 2.5
GLM 5
Mimimax 2
26tps 10000 Mac Studio last gen
M3 ultra
50tp 8k
Agents outside of coding
UI innovation (openclaw, codex desktop)
Major SLM Investment
The Real Jaseci Products
Jac-GPT - Interesting agent 2 -> (Open source, enterprise edition (Open source, dashboard configure integration, onprem tooling)-- Retool JacGPT -> AnyGPT
Dashboard/configurator/integration-- Desktop App, white-label customizable-- Auto specialize
-- JacGPT -> AnyGPT and Open source launch
-- Laucn MVP Enterprise edtion (autospecialize not required)
-- Autopecialize is priority in AnyGPT roadmap, we build accordingly
Autospecialize IP
Special Push button auto-specialize:
Step 1: use opus to get many question from docs
Step 2: Use opus to run agent on questions to generate dataset
Step 3: Use dataset to train SLM to be as good as opus on docs
Jac coder semi auto specialize
Step 1: Run product with opus for a while to get data
Step 2: Use dataset to train SLM to be as good as opus on docs
Kimi 2.5
GLM 5
Mimimax 2
26tps 10000 Mac Studio last gen
M3 ultra
50tp 8k
Agents outside of coding
UI innovation (openclaw, codex desktop)
Major SLM Investment