Amazon Web Services has made Kiro bots generally available. Announced on Thursday, November 5, 2026, they are named, persistent AI agents inside Kiro Crew, the team workspace for Amazon’s AI coding tool Kiro. Each bot owns one ongoing job, such as watching for failed builds or checking whether a release is ready, and keeps its own memory between runs.
What a Kiro bot does
According to VentureBeat’s report, a bot has a name, a continuing conversation, its own notes and a dedicated workspace. It can direct several worker sessions, so one bot can coordinate other agents on a task. Example jobs include investigating failures overnight, reviewing proposed code changes and producing weekly change summaries. Developers can describe schedules in plain language, and an activity log records routine work while the conversation surfaces findings and questions that need a person.
Memory is tied to the bot’s assignment, so a deployment bot’s decisions stay separate from a research bot’s context. Users can inspect, revise or remove what a bot remembers. AWS says bots improve through saved context and corrections, not by retraining the underlying models.
Permissions
Permissions are set per bot. A research bot can read files without editing them, a deployment bot asks for approval before making changes, and a review bot can comment and request revisions but cannot approve or merge code. Adding tools needs user approval, and the boundaries can be adjusted later. Crew can run on a laptop, a Mac Mini or an Amazon EC2 instance, and the coordinator and its workers can use different models.
What it costs
AWS has announced no separate per-bot fee. Bot work draws on the same credits as other Kiro use. Kiro’s pricing page lists these plans:
- Free: $0, 50 credits
- Pro: $20 a month, 1,000 credits
- Pro+: $40 a month, 2,000 credits
- Pro Max: $100 a month, 5,000 credits
- Power: $200 a month, 10,000 credits
- Teams (credit pooling): $20 per seat a month, two or more seats per pool, 500 credits per seat added to a shared pool
Extra credits cost $0.04 each. On individual plans they come in add-on packs from $5 for 125 credits up to $100, are billed at purchase and expire after 12 months. Unused monthly plan credits do not roll over. For teams, admins get alerts at 80 and 100 percent of pool usage, and the page says AWS GovCloud pricing is about 20 percent higher with no free tier.
The $20 figure that appears in launch coverage is therefore the existing team seat price, not a charge for bots. A team of five on pooled seats would pay $100 a month for 2,500 included credits, by our arithmetic from the page. Anything beyond that is billed at $0.04 per credit, so 1,000 extra credits would add $40.
The cost question to watch
A credit is a measure of work, not a finished task. The pricing page says simple prompts can cost under one credit while complex tasks such as spec execution typically cost more, and model choice changes the rate. A bot that runs on a schedule every night will therefore use credits at a rate that depends on how much it does and which model it uses. The subscription price alone does not tell a team what a continuously running bot will cost, and VentureBeat notes the announcement offers no benchmarks, productivity results or error rates.
Teams weighing bots should start with one narrowly scoped job, watch the credit meter for a few weeks and use the admin alerts before adding more. For another recent shift in how AI plans bill usage, see our report on the shared ChatGPT and Codex allowance.
Sources
- VentureBeat: AWS launches agentic AI Kiro bots with memory and support for recurring development jobs
- Kiro: Pricing
Like our coverage? Add stepphase.com as a preferred source on Google to see more of our stories in Top Stories.
