BoldCreate by Syncfusion

AI coding agent CLI for your terminal and CI

Pipe in the problem and get back a fix that passed its tests

Copying errors into a chat loses the context. Pipe a failing log into boldcreate and it reads the code, fixes it, and runs the tests. In a pipeline it reviews every branch and answers with an exit code.

macOS
curl -fsSL https://downloads.boldcreate.ai/cli/install.sh | bash
Windows
irm https://downloads.boldcreate.ai/cli/install.ps1 | iex
Read the CLI installation guide
Built by Syncfusion, serving 47,000+ organizations.25+ years of developer tools, now reimagined for the agent-first era.
47,000+Syncfusion customers
1,600+Syncfusion UI components
25+Years of Syncfusion developer tools
Syncfusion component licensing is separate.
Three ways in

Three ways to run the CLI

boldcreateInteractive sessionAn interactive session in this folder. Plan, edit, and test together, with every command shown first.boldcreate --agent plan "split OrderService"
boldcreate -pAsk once, then exitOne answer and done. Pipe logs, diffs, or stack traces straight in with -p -.git diff | boldcreate -p - "anything risky?"
boldcreate -cContinue the last sessionContinue the last session in this folder, context and approved plan included.boldcreate sessions # choose another
In your pipeline

Use CLI results in a CI step

Sign in with an organization key, run a prompt with --mode json, and branch on the result. Exit 3 holds the build until a person approves in BoldCreate.

# .github/workflows/review.yml- run: npm ci && npm run build- run: boldcreate login --api-key $BC_ORG_KEY- run: | boldcreate -p "review this branch" \ --mode json --model auto \ --no-session > review.json echo "exit $?"
0PassedNo findings. Deploy continues.
1FailedFindings break your rules. Report attached.
2Usage errorBad flag or missing key. Fix the step.
3Needs approvalPaused until a person approves.
--model auto

Select models to suit each review job

CI runs hundreds of times a day. With Auto, routine reads and checks go to a fast, low-cost model and only real findings get the strong one. Pin a model with --model when a job needs it.

Illustrative example, not a quoteCI reviews this month0 runs
On Auto$0.00
Each bar shows one day of CLI reviews on Auto
Four machines, one policy

Check shared rules from any CLI session

Laptop, CI runner, build server over SSH, or a teammate’s terminal: boldcreate status shows the same skills, approval policy, and budget, because none of it lives on the machine.

See how plugins add skills and rules

Frequently asked questions

How do I install the AI coding agent CLI?
On macOS, run curl -fsSL https://downloads.boldcreate.ai/cli/install.sh | bash. On Windows, run irm https://downloads.boldcreate.ai/cli/install.ps1 | iex in PowerShell. Then run boldcreate in any folder to start an interactive session where you plan, edit, and test together, with every command shown first.
What modes does the CLI support?
Run boldcreate to work alongside it, boldcreate -p to ask once and exit, or boldcreate -c to continue the last session in a folder with its context and approved plan. Pipe logs, diffs, or stack traces straight in with -p -.
How does the CLI work in a CI pipeline?
Sign in with an organization key, run a prompt with --mode json, and branch on the result. Exit 0 means passed, 1 failed, 2 a usage error, and 3 holds the build until a person approves in BoldCreate.
Will frequent reviews get expensive?
Auto keeps routine reads and checks on a fast, low-cost model and saves the strong one for real findings. Pin a model with --model when a job needs it.
Where do the CLI’s rules and budgets live?
In your organization, not in a dotfile. On a laptop, a CI runner, a build server over SSH, or a teammate’s terminal, boldcreate status shows the same skills, approval policy, and budget.
Which Git hosts does BoldCreate work with?
BoldCreate works with repositories hosted on GitHub, GitLab, and Gitea.

Add a review step to one pipeline

One organization key, one prompt, four exit codes. Try it on a branch today.

$199 a month for 20 seats · AI and sandboxes at provider cost + 10% platform fee