You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
docs(blog): add Langdock vs Dust 2026 comparison article - #8136
Adds the head-to-head comparison article Langdock vs Dust: Enterprise AI Workspace, Agents, and Pricing Compared (2026) (/blog/langdock-vs-dust) along with its 1600x900 blog cover, 2 custom 1600x900 architectural and pricing diagrams, and 3 1600px-wide .webp product screenshots.
Key Highlights
Architectural comparison: Compares Langdock's EU-governed multi-model employee chat workspace, custom Agents, Skills, Canvas apps, and separate Workflows builder (2,500 to 10,000 included runs/month or +€449/month billed annually for the 40,000-run Workflows Business add-on with separate token billing) against Dust's (dust.tt) permissioned Knowledge Spaces (Open, Restricted, Project), multi-agent @mention orchestration in Pods, warehouse SQL querying, and interactive React Frames (viz).
2026 pricing & scaling limits verified from primary sources: Documents Langdock's rolling 5-hour session and Monday-to-Monday weekly windows (throttling to GPT-6 Luna at 100%), Core Governance unbundling on Jan 1, 2027 (+€3.50/user/month), and 2,000+ / 5,000+ seat thresholds for single-tenant cloud and self-hosted Kubernetes; alongside Dust's Business plan (€24/seat/month Pro and €120/seat/month Max billed annually, up to 100 seats) caps of 3 managed data-source connectors, 5 Spaces, and 5 remote MCP servers.
Open-source AI coworker alternative (Agenta): Details how Agenta (MIT-licensed) removes per-seat pricing cliffs and connector caps by running real coding-agent harnesses (Claude Code, Codex, Pi) inside isolated Daytona Linux and Chromium sandboxes with persistent shared agent files, 1,000+ integrations, tool-level approvals (allow/ask/deny), and multi-surface publishing (Slack, Telegram, WhatsApp, and embeddable web chat widgets), self-hosted free from 1 user up or on Agenta Cloud with unlimited users ($0 Hobby, $29/mo Starter, $299/mo Business).
Added a 2026 comparison of Langdock, Dust, and Agenta, covering their features, pricing, hosting options, usage limits, and integrations.
Includes comparison tables, example cost scenarios, platform-selection guidance, and FAQs.
Adds a 2026 article comparing Langdock, Dust, and Agenta. It covers their capabilities, usage and plan limits, pricing, compliance, hosting, and deployment options. It also includes cost scenarios, platform-selection guidance, and FAQs.
Changes: Platform comparison article
Layer / File(s)
Summary
Platform comparison and guidance web/website/src/content/posts/langdock-vs-dust.mdx
Adds comparisons of platform architecture, capabilities, limits, pricing, compliance, hosting, and deployment. Includes cost scenarios, selection guidance, and FAQs.
The comparison article contains factual and pricing inaccuracies that could mislead readers choosing a platform, particularly around Agenta's free-plan team limit and the Scenario A cost. Fix these before publishing; there is no runtime or site-functionality risk.
Pre-merge checks: 5
Check name
Status
Explanation
Title check
The title clearly identifies the Langdock vs Dust comparison article.
Description check
The description matches the comparison article and its coverage of Langdock, Dust, and Agenta.
Docstring Coverage
No eligible functions found in the changed files. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documentat…
Linked Issues check
Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check
Check skipped because no linked issues were found for this pull request.
Full details: Docstring Coverage
Explanation
No eligible functions found in the changed files. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documentation-only edits and deletions within surviving functions are included; unchanged diff context and fully deleted functions are excluded. Analyzed 0 functions across 0 files. (1 skipped: 1 unsupported.)
Finishing Touches
Generate unit tests (beta)
Commit to this branch
Create a new PR
Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.
Included review availability: This review used your included allowance. Your plan provides up to 8 included reviews per hour; 7 remain after this review.
- **Langdock** starts from the **compliance-first multi-model employee workspace**. Headquartered in Berlin, it gives every employee a governed chat interface, prompt library, reusable **Skills**, meeting note transcription, and custom **Agents**, while keeping background automations inside a separate visual **Workflows** product (`2,500` to `10,000` included runs per month, or `€449/month` billed annually for the `40,000`-run Workflows Business add-on where model tokens are billed separately). Its pricing (`€23.20/user/month` on Core or `€31.20/user/month` on Work for the first 50 standard seats) relies on rolling **5-hour session windows and Monday-to-Monday weekly windows**, and its self-hosted Kubernetes option requires at least **5,000 seats**.
- **Dust (`dust.tt`)** starts from **company-knowledge agents and collaborative workspaces**. Founded in Paris and San Francisco, Dust organizes internal knowledge into permissioned **Spaces** (`Open`, `Restricted`, and `Project` Spaces), lets teammates chain specialized `@agents` inside shared **Pods**, and renders live interactive React apps and dashboards called **Frames** that refresh on wake-up schedules. Yet its **Business plan** (`€24/seat/month` Pro or `€120/seat/month` Max billed annually, up to 100 seats) caps the entire workspace at **3 managed data-source connectors, 5 Spaces, and 5 remote MCP servers** until you upgrade to a 100+ seat Enterprise contract.
- **Neither Langdock nor Dust is open source, and neither gives every agent a real coding-agent harness or full Linux workstation.** For teams that want plain-language agent building for non-technical teammates without per-seat pricing cliffs, connector caps, or 5,000-seat self-hosting minimums, **[Agenta](https://agenta.ai)** provides a third architectural path: an [MIT-licensed](https://opensource.org/license/mit) open-source AI coworker platform where every agent runs on a battle-tested coding agent harness ([Claude Code](https://code.claude.com/docs/en/overview), [OpenAI Codex](https://developers.openai.com/codex), or [Pi](https://github.com/badlogic/pi-mono/)) inside an isolated [Daytona](https://www.daytona.io/) Linux and browser sandbox.
Agenta agents sandboxes and security local provider Daytona self-hosted docs
Result:
Sources consulted:
- Sandboxes and security - Docs - Agenta — https://agenta.ai/docs/reference/agents/sandboxes-and-security
- Run agents in a cloud sandbox (Daytona) - Docs - Agenta — https://agenta.ai/docs/self-host/guides/agent-daytona-sandboxes/
- agenta-skills/skills/self-host-agenta/SKILL.md at main · Agenta-AI/agenta-skills · GitHub — https://github.com/Agenta-AI/agenta-skills/blob/main/skills/self-host-agenta/SKILL.md
- Quick Start: Self-Hosting Agenta - Docs - Agenta — https://agenta.ai/docs/self-host/host-locally/
- …and 22 more
Distinguish Daytona from self-hosted local execution.
Not every Agenta agent runs in Daytona. Agenta Cloud uses Daytona, but self-hosted deployments can use the local provider or configure Daytona separately. Apply this distinction to the other architecture descriptions and FAQs. (Agenta sandbox documentation)
| **Agent & Workflow Architecture** | Chat **Agents** + reusable **Skills** + separate trigger-action **Workflows** module (up to `2,000` steps/run) | Multi-agent `@mention` orchestration, sub-agent tool calls, webhook/schedule triggers, and wake-up **Frames** | Real coding agent harnesses ([Claude Code](https://code.claude.com/docs/en/overview), [Codex](https://developers.openai.com/codex), [Pi](https://github.com/badlogic/pi-mono/)) with cron schedules, event subscriptions, and tool approvals |
| **Code & File Execution** | Sandboxed Python data analysis and interactive Canvas web apps (`5-user` share cap on Core; unlimited on Work) | Interactive React **Frames** (`viz`) and SQL warehouse querying (BigQuery, Snowflake, Salesforce); no Linux VM | Dedicated [Daytona](https://www.daytona.io/) Linux VM and headless Chromium browser per agent + persistent shared agent files and session scratch files |
| **Integrations & MCP** | `57` native integrations (`754` actions), custom OpenAPI actions, and remote MCP server support | Managed continuous knowledge sync + remote MCP servers (**Business plan capped at `3` connectors and `5` MCP servers**) | `1,000+` managed OAuth integrations, custom MCP servers, per-tool `allow` / `ask` / `deny` gates, and full CLI (`git`, `gh`, `curl`) |
Agenta pricing Hobby Starter Business team members unlimited
Result:
Sources consulted:
- Plans and credits - Docs - Agenta — https://agenta.ai/docs/administration/billing/plans-and-credits
- Pricing — Agenta — https://www.agenta.ai/pricing
- Credit-based plans and reliability fixes - Docs - Agenta — https://agenta.ai/docs/changelog/credit-plans-and-reliability-fixes
- 7 Best Open-Source AI Agent Platforms in 2026 (Compared) — Agenta Blog — https://www.agenta.ai/blog/best-open-source-ai-agent-platforms
- 9 Best Agentic AI Tools in 2026 (For Teams, Workflows, and Developers) — Agenta Blog — https://agenta.ai/blog/best-agentic-ai-tools
- …and 14 more
Correct the Cloud Hobby team limit.
Hobby permits two team members. Lines 199 and 216 incorrectly say it has unlimited seats. Correct both claims, and make the limit explicit in the table and FAQ.
Suggested fix for the unlimited-seat claims
-6. **MIT Open-Source Self-Hosting and Unlimited-Seat Cloud Pricing:** Self-host Agenta free forever via [Docker Compose](https://docs.docker.com/compose/) or [Kubernetes Helm](https://helm.sh/) with **unlimited users, SSO, and RBAC included in the free open-source edition**. Or run on Agenta Cloud with **unlimited seats on every plan**: **Hobby at $0 forever** (`500` signup bonus credits + `75` daily credits on up to 10 days/month), **Starter at $29/month** (`2,900` credits), and **Business at $299/month** (`32,000` credits, SSO, SOC 2 Type II), with **$0 Agenta credit charges on model tokens when you bring your own API keys or ChatGPT subscription**.+6. **MIT Open-Source Self-Hosting and Unlimited-Seat Cloud Pricing:** Self-host Agenta free forever via [Docker Compose](https://docs.docker.com/compose/) or [Kubernetes Helm](https://helm.sh/) with **unlimited users, SSO, and RBAC included in the free open-source edition**. Or run on Agenta Cloud with **two team members on Hobby and unlimited seats on Starter and Business**: **Hobby at $0 forever** (`500` signup bonus credits + `75` daily credits on up to 10 days/month), **Starter at $29/month** (`2,900` credits), and **Business at $299/month** (`32,000` credits, SSO, SOC 2 Type II), with **$0 Agenta credit charges on model tokens when you bring your own API keys or ChatGPT subscription**.
-- You want **MIT-licensed open-source self-hosting from 1 user up** (with SSO and RBAC included free) or **unlimited-user cloud pricing** (`$0` Hobby / `$29/mo` Starter / `$299/mo` Business) without paying a `€24` to `€150` monthly tax for every teammate you invite.+- You want **MIT-licensed open-source self-hosting from 1 user up** (with SSO and RBAC included free) or cloud pricing with two team members on `$0` Hobby and unlimited team members on `$29/mo` Starter and `$299/mo` Business.
| **Scenario A: 30-person team using daily AI chat, 5 SaaS knowledge sources, and light automations** | **€696/mo** on Core (`30 × €23.20`) or **€936/mo** on Work (`30 × €31.20`), plus `€105/mo` starting Jan 2027 if Core needs Governance | **Blocked on Business by the 3-connector cap** (`€720/mo` for 30 Pro seats only connects 3 data sources; connecting 5 managed sources requires an Enterprise contract) | **$0 software license** self-hosted, or **$29/mo** (Starter) to **$299/mo** (Business) on Agenta Cloud for **all 30 users** (unlimited seats, zero per-seat fees) |
The reason will be displayed to describe this comment to others. Learn more.
Minor · Functional Correctness · Quick win
Supported by static analysis
Script executed:
sed -n 160,170p web/website/src/content/posts/langdock-vs-dust.mdx
rg -n -i 'credit' web/website/src/content/posts/langdock-vs-dust.mdx
Repository: Agenta-AI/agenta
Length of output: 8417
Web query:
Agenta plans and credits Starter 2,900 credits BYOK model credit markup
Result:
Sources consulted:
- Plans and credits - Docs - Agenta — https://agenta.ai/docs/administration/billing/plans-and-credits
- Credit-based plans and reliability fixes - Docs - Agenta — https://agenta.ai/docs/changelog/credit-plans-and-reliability-fixes
- Pricing — Agenta — https://agenta.ai/pricing
- 9 Best Agentic AI Tools in 2026 (For Teams, Workflows, and Developers) — Agenta Blog — https://agenta.ai/blog/best-agentic-ai-tools
- Agenta Features & Pricing: 12.5/14 Score | Agentic Index — https://agenticindex.io/vendors/agenta
- …and 13 more
Identify the Cloud prices as base subscriptions.
Starter’s credits may not cover Scenario A’s daily use. BYOK removes model credit charges, but runtime still uses credits. Additional credits may be needed.
Suggested fix
--- "a/web/website/src/content/posts/langdock-vs-dust.mdx"+++ "b/web/website/src/content/posts/langdock-vs-dust.mdx"@@ -161,7 +161,7 @@
| Team Scenario | Langdock Monthly Cost | Dust (`dust.tt`) Monthly Cost | Agenta (MIT Self-Hosted or Cloud) |
| :--------------------------------------------------------------------------------------------------------------- | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------- | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------- | :---------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| **Scenario A: 30-person team using daily AI chat, 5 SaaS knowledge sources, and light automations** | **€696/mo** on Core (`30 × €23.20`) or **€936/mo** on Work (`30 × €31.20`), plus `€105/mo` starting Jan 2027 if Core needs Governance | **Blocked on Business by the 3-connector cap** (`€720/mo` for 30 Pro seats only connects 3 data sources; connecting 5 managed sources requires an Enterprise contract) | **$0 software license** self-hosted, or **$29/mo** (Starter) to **$299/mo** (Business) on Agenta Cloud for **all 30 users** (unlimited seats, zero per-seat fees) |+| **Scenario A: 30-person team using daily AI chat, 5 SaaS knowledge sources, and light automations** | **€696/mo** on Core (`30 × €23.20`) or **€936/mo** on Work (`30 × €31.20`), plus `€105/mo` starting Jan 2027 if Core needs Governance | **Blocked on Business by the 3-connector cap** (`€720/mo` for 30 Pro seats only connects 3 data sources; connecting 5 managed sources requires an Enterprise contract) | **$0 software license** self-hosted, or a **base subscription of $29/mo** (Starter) to **$299/mo** (Business) on Agenta Cloud for **all 30 users** (unlimited seats, zero per-seat fees); additional credits may be needed |
| **Scenario B: 50-person company with 10 heavy agent builders and 35,000 monthly background workflow runs** | **€1,160 to €2,369/mo+** (`40` Core seats at `€23.20` + `10` 5x seats at `€99` = `€1,918/mo`, plus **€449/mo** Workflows Business add-on + separate API token charges) | **€2,160/mo** billed annually (`40` Pro seats at `€24` + `10` Max seats at `€120`, still capped at `3` managed connectors and `5` MCP servers on Business) | **$0 software license** self-hosted + raw BYOK LLM tokens, or **$299/mo** on Agenta Cloud Business with unlimited users and **$0 model credit markup on BYOK** |
| **Scenario C: 40-person regulated engineering & ops org requiring self-hosted Docker/Kubernetes, SSO, and RBAC** | **Not available** (Langdock requires **2,000+ seats** for single-tenant cloud and **5,000+ seats** for self-hosted Kubernetes) | **Not available** (Dust is cloud-only SaaS with no self-hosted edition; single-tenant cloud requires **100+ seat** Enterprise) | **$0 forever** on **Agenta Open Source (MIT)** via Docker Compose or Helm on Kubernetes, with **SSO, RBAC, and unlimited users included free** |
Committable suggestion
IMPORTANT
Review the replacement and indentation, then test before committing.
Suggested change
|**Scenario A: 30-person team using daily AI chat, 5 SaaS knowledge sources, and light automations**|**€696/mo** on Core (`30 × €23.20`) or **€936/mo** on Work (`30 × €31.20`), plus `€105/mo` starting Jan 2027 if Core needs Governance |**Blocked on Business by the 3-connector cap** (`€720/mo` for 30 Pro seats only connects 3 data sources; connecting 5 managed sources requires an Enterprise contract) |**$0 software license** self-hosted, or **$29/mo** (Starter) to **$299/mo** (Business) on Agenta Cloud for **all 30 users** (unlimited seats, zero per-seat fees) |
|**Scenario A: 30-person team using daily AI chat, 5 SaaS knowledge sources, and light automations**|**€696/mo** on Core (`30 × €23.20`) or **€936/mo** on Work (`30 × €31.20`), plus `€105/mo` starting Jan 2027 if Core needs Governance |**Blocked on Business by the 3-connector cap** (`€720/mo` for 30 Pro seats only connects 3 data sources; connecting 5 managed sources requires an Enterprise contract) |**$0 software license** self-hosted, or a **base subscription of $29/mo** (Starter) to **$299/mo** (Business) on Agenta Cloud for **all 30 users** (unlimited seats, zero per-seat fees); additional credits may be needed|
| **Scenario A: 30-person team using daily AI chat, 5 SaaS knowledge sources, and light automations** | **€696/mo** on Core (`30 × €23.20`) or **€936/mo** on Work (`30 × €31.20`), plus `€105/mo` starting Jan 2027 if Core needs Governance | **Blocked on Business by the 3-connector cap** (`€720/mo` for 30 Pro seats only connects 3 data sources; connecting 5 managed sources requires an Enterprise contract) | **$0 software license** self-hosted, or **$29/mo** (Starter) to **$299/mo** (Business) on Agenta Cloud for **all 30 users** (unlimited seats, zero per-seat fees) |
| **Scenario B: 50-person company with 10 heavy agent builders and 35,000 monthly background workflow runs** | **€1,160 to €2,369/mo+** (`40` Core seats at `€23.20` + `10` 5x seats at `€99` = `€1,918/mo`, plus **€449/mo** Workflows Business add-on + separate API token charges) | **€2,160/mo** billed annually (`40` Pro seats at `€24` + `10` Max seats at `€120`, still capped at `3` managed connectors and `5` MCP servers on Business) | **$0 software license** self-hosted + raw BYOK LLM tokens, or **$299/mo** on Agenta Cloud Business with unlimited users and **$0 model credit markup on BYOK** |
The reason will be displayed to describe this comment to others. Learn more.
Minor · Functional Correctness · Quick win
Recalculate Scenario B’s Langdock range.
The stated seat mix costs 40 × €23.20 + 10 × €99 = €1,918. Adding the required €449 Workflows Business package gives €2,367/month before tokens, not €2,369. The €1,160 lower bound excludes both the ten specified heavy-use seats and the package needed for 35,000 runs. Replace the range with the cost of the stated configuration. (langdock.com)
Dependent tests finished. Post /preview for manual review. Environment absence verified; retained storage requires separate provider accounting.
Post /preview as a new comment to request a one-hour preview. Repeating it while active does not extend the hour.
Preview data is disposable. Expiry is enforced by the next successful cleanup sweep, not an exact timer.
Environment deletion stops its compute. Provider-retained storage and permanent services may still be billed.
This branch has not been deployed
No deployments
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds the head-to-head comparison article Langdock vs Dust: Enterprise AI Workspace, Agents, and Pricing Compared (2026) (
/blog/langdock-vs-dust) along with its1600x900blog cover, 2 custom1600x900architectural and pricing diagrams, and 31600px-wide.webpproduct screenshots.Key Highlights
2,500to10,000included runs/month or+€449/monthbilled annually for the40,000-run Workflows Business add-on with separate token billing) against Dust's (dust.tt) permissioned Knowledge Spaces (Open,Restricted,Project), multi-agent@mentionorchestration in Pods, warehouse SQL querying, and interactive React Frames (viz).GPT-6 Lunaat 100%), Core Governance unbundling on Jan 1, 2027 (+€3.50/user/month), and2,000+/5,000+seat thresholds for single-tenant cloud and self-hosted Kubernetes; alongside Dust's Business plan (€24/seat/monthPro and€120/seat/monthMax billed annually, up to 100 seats) caps of3managed data-source connectors,5Spaces, and5remote MCP servers.Claude Code,Codex,Pi) inside isolated Daytona Linux and Chromium sandboxes with persistent shared agent files, 1,000+ integrations, tool-level approvals (allow/ask/deny), and multi-surface publishing (Slack, Telegram, WhatsApp, and embeddable web chat widgets), self-hosted free from 1 user up or on Agenta Cloud with unlimited users ($0Hobby,$29/moStarter,$299/moBusiness).Assets Added
web/website/src/content/posts/langdock-vs-dust.mdxweb/website/public/blog/langdock-vs-dust/hero.webp(1600x900)web/website/public/blog/langdock-vs-dust/langdock-vs-dust-architecture.webp(1600x900)web/website/public/blog/langdock-vs-dust/langdock-vs-dust-pricing-mechanics.webp(1600x900)web/website/public/blog/langdock-vs-dust/langdock.webp(1600x922)web/website/public/blog/langdock-vs-dust/dust.webp(1600x822)web/website/public/blog/langdock-vs-dust/agenta.webp(1600x995)