camelAI
Also known as: camelCode
camelAI sells camelCode, an MIT licensed coding agent that builds projects in a persistent workspace and publishes them to live URLs, hosted or self hosted on the customer's own model keys, alongside camelStream, a flat rate inference service.
camelAI began as an AI data analyst and now sells two products. camelCode is a coding agent: a user describes what they want and it builds the project in a workspace whose files, databases and deployed apps persist across sessions, runs builds, notebooks and database queries in isolated containers, and publishes the result to a live URL. It connects to PostgreSQL, MongoDB, BigQuery, Supabase and ClickHouse, and paid plans add cron jobs, custom domains and an email inbox.
camelCode is open source under the MIT license and can be self hosted with Docker Compose or on AWS behind Pomerium authentication with the customer's own identity provider, running on a key for Bedrock, Anthropic, OpenAI, OpenRouter or a compatible provider; hosted plans also take the customer's own key. Self hosted deployments add spend limits and usage tracking.
camelStream, sold beside it, is a flat rate inference service at $5 a month that serves a changing fleet of models through OpenAI and Anthropic compatible APIs and powers camelCode's free tier; the grid here grades camelCode, not camelStream. camelCode has a free plan, with paid plans from $10 a month.
Vendor details
Canonical URL
https://camelai.com
Category
Coding agent
Funding status
Independent. Founding team previously shipped several products together. Specific funding figures were not retrieved this session.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Connects to 50+ tools and data sources out of the box, including PostgreSQL, MySQL, ClickHouse, DuckDB, BigQuery, Snowflake, Supabase, MongoDB, plus Stripe, HubSpot, Salesforce, Notion, and GitHub. Talk to the agent via chat, its own workspace email address, or Slack.
Sources & related URLs
Research sources
Agentic Index coverage score
6.0 / 14 capabilities · 43%
| Integrations & Tool Calling | Partial |
|---|---|
|
camelCode, the coding agent camelAI now sells, connects to PostgreSQL, MongoDB, BigQuery, Supabase and ClickHouse and runs database queries, builds and notebooks in isolated containers, then publishes the apps it builds to live URLs on camelAI's own hosting; no action in other companies' systems and no custom tool mechanism is documented. Sourcecamelai.comread 2026-09-28 |
|
| Workflow Orchestration | Partial |
|
camelCode takes a description and builds, runs and publishes the project in one agent loop, with cron jobs on the paid plans; no branching, conditions, multiple cooperating agents or flow the buyer composes is documented, and the July multi worker deep research agent is no longer on the site. Sourcecamelai.comread 2026-09-28 |
|
| Knowledge Grounding & RAG | Not documented |
|
No knowledge base, retrieval index or semantic layer over the customer's documents or data is described for camelCode; the July knowledge base and reference query system belonged to the data analyst product that is no longer on the site. Sourcecamelai.comread 2026-09-28 |
|
| Human Oversight & Guardrails | Partial |
|
Self hosted camelCode documents spend limits and usage tracking, customer set constraints on what the agent can consume, and runs builds, notebooks and queries in isolated containers; no approval step before the agent publishes an app or runs a query is documented. Sourcecamelai.com/docs/self-hosting/overviewread 2026-09-28 |
|
| Security, Identity & Governance | Partial |
|
Self hosted camelCode sits behind Pomerium authentication with OIDC to the customer's own identity provider, and a Team plan is priced per seat; camelAI points to a trust center at trust.oneleet.com/camelai, though no attestation from it is published in readable form, and the CASA and SOC 2 in progress claims made in July no longer appear on the site. Sourcecamelai.com/docs/self-hosting/overviewread 2026-09-28 |
|
| Observability & Auditability | Partial |
|
Each project keeps its source and history in an artifact repository and self hosted deployments document usage tracking and an operations guide, but no trace of the agent's steps, tool calls or commands per run is described. Sourcecamelai.com/docs/self-hosting/overviewread 2026-09-28 |
|
| Memory & State Persistence | Partial |
|
camelCode works in a workspace whose files, databases and deployed apps survive every session, kept in durable storage, which is a workspace that keeps its own state; no memory store the agent reads as context with a stated scope and lifetime is documented. Sourcecamelai.comread 2026-09-28 |
|
| Deployment & Data Residency | Full |
|
Besides the hosted service, camelCode is MIT licensed and self hosts on the customer's own infrastructure with Docker Compose or on AWS, keeping application state, project storage and build, notebook and query containers local, with the stated limits that self hosted camelAI sends no email and has no multi node failover. SourcecamelAI, docs Self-host Coding Agent overview and github.com/qaml-ai/camelAI README; camelai.com/docs/self-hosting/overviewread 2026-09-28 |
|
| Prebuilt Agents / Templates / Packs | Not documented |
|
camelCode builds whatever the user describes; no templates, starter projects or prebuilt agents are offered. Sourcecamelai.comread 2026-09-28 |
|
| Triggers & Channel Coverage | Partial |
|
Plans include cron jobs (daily on Free, hourly on Starter, every five minutes on Pro) and an email inbox on Starter, but whether a cron job or an inbound email wakes the agent or only runs the apps it built is not stated; otherwise the user starts the agent by describing a task. Sourcecamelai.com/pricingread 2026-09-28 |
|
| Model Flexibility & Routing | Full |
|
Every hosted camelCode plan takes the customer's own API key, and self hosted camelCode runs on a key the customer supplies for Bedrock, Anthropic, OpenAI, OpenRouter or a compatible custom provider; the free tier runs on camelAI's camelStream fleet, which camelAI routes. Sourcecamelai.com/docs/self-hosting/overviewread 2026-09-28 |
|
| APIs / SDKs / MCP Extensibility | Partial |
|
camelCode's source is public under MIT and the self hosting docs include a guide to customizing the coding agent, but no API or SDK for driving camelCode is documented; the documented OpenAI and Anthropic compatible API belongs to camelStream, the inference service sold beside it. SourcecamelAI, github.com/qaml-ai/camelAI README and docs indexread 2026-09-28 |
|
| Testing, Debugging & Optimization | Not documented |
|
No way to test, score or compare the agent's work is described; the repository's own test commands guard camelAI's product, not the customer's agent runs. SourcecamelAI, homepage, docs and github.com/qaml-ai/camelAI README; camelai.comread 2026-09-28 |
|
| Browser / Computer-use | Not documented |
|
camelCode executes code in containers and a persistent workspace, which is code execution rather than control of a browser, desktop or computer interface; no browser or GUI control is described. Sourcecamelai.com/docs/self-hosting/overviewread 2026-09-28 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Recent platform changes
camelAI launched camelRun, an open source hosted runtime for long lived agents that runs the agent loop, keeps each agent's history and files, runs model written code in a sandbox with no network access, and wakes sleeping agents when work arrives. It costs $0.01 per hour of active agent time plus model usage, and can be self hosted with one container and Postgres.
Bears on: Memory / state
View sourcecamelAI completely redesigned its coding agent architecture, migrating from per-user virtual machines to a serverless model using Cloudflare Durable Objects. The updated execution stack stores project files in SQLite and R2, replacing open-ended bash commands with sandboxed JavaScript Code Mode running in V8 isolates.
Bears on: Agent capability
View sourcecamelAI introduced a free tier powered by a self-hosted DeepSeek V4 Flash model running on AWS spot instances. The architecture routes requests through a Cloudflare AI Gateway with automatic failover to Azure's hosted DeepSeek service.
Bears on: Pricing / packaging
View sourcePricing
camelCode free plan; Starter $10 per month (camelStream inference, sold separately, $5 per month)
per plan per month with included model credits, usage after credits; Team per seat
What is public
camelCode Free, Starter, Pro and Team prices with their limits, and camelStream at $5 a month.
Cost watchouts
Model usage beyond the included credits adds to the plan price unless the customer brings its own key.
Variable cost rationale
Combines seats with query and agent run usage; cost grows with data volume and scheduled agent activity, though a free tier caps entry cost.
Overage / add-ons
Usage beyond the plan's included model credits is billed; every plan can use the customer's own API key
Sales call required
No, self serve available
Free / trial
camelCode Free plan, no card; bring your own key
Lowest paid plan
camelCode Starter ($10 per month)
Key ambiguities
The Team per seat figure appears in the page metadata rather than a plan card; how usage after credits is priced is not stated.
Related vendors
- Cognition — Maker of Devin, an autonomous AI software engineer
- 10Web — Agentic website platform whose specialized AI agents build, host,…
- AgentUI — Managed AI app builder for operations teams that generates and hosts…
- Aider — Open source, model agnostic terminal coding agent that edits your…
- Anthropic Claude Code — Anthropic's agentic coding system across terminal, desktop, IDE, web…
- AppFactor — Agentic platform that runs persistent agents across an enterprise…
Alternatives to camelAI
The closest documented capability profiles to camelAI among coding agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Tabby6.0 / 14Adds documented Knowledge Grounding & RAG
- Blackbox AI7.5 / 14Adds documented Knowledge Grounding & RAG and Testing, Debugging & Optimization
- iGent5.0 / 14Adds documented Knowledge Grounding & RAG and Testing, Debugging & Optimization
- Cosine8.5 / 14Adds documented Knowledge Grounding & RAG and Prebuilt Agents, Templates & Packs, among others
- Charm10.0 / 14Adds documented Knowledge Grounding & RAG and Prebuilt Agents, Templates & Packs, among others
- GoCodeo7.0 / 14Adds documented Knowledge Grounding & RAG and Prebuilt Agents, Templates & Packs, among others
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded