Best Self-Hosted Feedback Tools (2026)
Among the best self-hosted feedback tools in 2026, Quell is the strongest pick for teams shipping with coding agents: managed cloud at £3/$4/€4 per seat in early access, and an MCP server that lets your agents read, comment, and triage feedback directly. If self-hosting isn’t a requirement and your team has no automation workflow, Userback or Marker.io will serve you well at comparable cost.
TL;DR — which to pick
Teams using coding agents (Cursor, Claude Code, etc.), anyone who wants self-host, or projects where a no-token-tax flat rate matters.
Small visual QA teams on a budget who don't need API access and are fine with cloud-only.
Agencies managing multiple client projects with a clean client-facing UX.
Non-technical stakeholders doing website review; simplest possible sticky-note UI.
Roadmap voting and feature-request boards. Not for visual bug capture with console/network context.
Idea and roadmap boards (MIT, self-hostable). Same category as Astuto — voting, not bug capture.
Feature & pricing
Six tools, side by side.
| Tool | Entry price | API / agent access | Self-host? | Best for |
|---|---|---|---|---|
| Quellbest fit | Cloud from £3/$4/€4/seat (early access) | Full REST API + CLI + MCP server (all tiers) | Roadmap (self-host) | Agent-native teams, self-host requirement |
| Userback | Free (2 seats) / $7/seat/mo (Team) | REST API on Business Plus only ($23/seat/mo) | No | Small visual QA teams, tight budget |
| Marker.io | $39/mo (3 seats) | REST API + integrations (paid plans) | No | Agencies, multi-client review |
| BugHerd | $42/mo (5 members) | API + integrations (paid plans) | No | Non-technical stakeholders, website review |
| Astuto | Free (open-source, AGPL) | Basic API | Yes (Docker) | Feature-voting boards |
| Fider | Free (open-source, MIT) | REST API | Yes (Docker) | Idea / roadmap boards |
Prices verified June 2026. Check vendor sites before relying on them. Marker.io and BugHerd expose APIs/integrations on paid plans; confirm the exact tier on their pricing pages.
Tool-by-tool
The full breakdown.
Quell
agent-nativeQuell is an agent-native feedback platform built on the premise that the people triaging your bugs are increasingly coding agents, not just humans. The widget captures bug reports with automatic context — screenshot with annotation, console logs, network calls, and page state — so reporters don’t describe their environment manually.
The same REST API that powers the admin inbox is open to agents via scoped keys, a CLI, and an MCP server (quell mcp, 10 tools). A trust-aware routing model controls what auto-routes vs. what needs sign-off. Agent-opened PRs are always review-required. Self-host is on the roadmap; cloud early access from $4/seat.
Quell is pre-product — no external paying users yet. The most credible proof is that it runs on itself: Quell’s own development uses Quell for feedback. If you need a tool backed by years of production use at scale, read the alternatives below first.
Userback
cloud-onlyA mature, cloud-hosted visual feedback tool. Reporters click a launcher, draw on the page, and the submission lands in a shared inbox with browser metadata. The free tier (two seats) is genuinely useful. The catch: REST API access is gated to Business Plus at $23/seat/mo. No self-host option.
Your team is small, your review process is human-driven, and you don't need API or agent access. The free tier is a real product, not a lead magnet.
Marker.io
agenciesTargets agencies and client-facing QA. Reporters annotate pages and submissions sync to Jira, ClickUp, GitHub, or Linear. Guest review is well thought out — clients submit without an account. Pricing is seat-bundled: $39/mo for 3 seats, $149/mo for 15. Cloud-only.
You run an agency with multiple client projects, need a clean guest-review UX, and your team is large enough that bundled-seat pricing works in your favour.
BugHerd
simplest UIThe simplest entry here. Stakeholders click anywhere to pin a sticky note; that note becomes a task with a screenshot and basic browser info. Zero training for guests. The trade-off is depth — less automatic context, designed for synchronous review rather than ongoing monitoring. Cloud-only.
Your reviewers are non-technical (clients, designers, copywriters) and simplicity of the submission experience outweighs depth of captured context.
Astuto & Fider
open sourceBoth open-source (AGPL / MIT), self-hostable, free, and run in Docker. They are feature-voting and roadmap boards: users submit ideas, upvote, comment. Neither captures console logs, network calls, or screenshots as part of a bug report — a different category of tool.
You want a public feature-request board, you're committed to open source, and you don't need deep bug-capture context. Budget: free beyond hosting.
FAQ
Straight answers.
What is the best self-hosted bug reporting tool?
Quell is currently the strongest pick in this category for combining visual bug capture, automatic console and network context, and an agent-accessible REST API. Astuto and Fider are self-hostable but are idea-voting boards, not bug capture tools. Self-host is on Quell's roadmap; cloud is $4/seat in early access.
Can I use a feedback tool with coding agents like Claude Code or Cursor?
Quell is purpose-built for this. Its MCP server (quell mcp) exposes 10 tools that let agents read feedback in Markdown, comment, update status, and open GitHub or Linear issues. It uses trust-aware routing so agent actions that require human oversight stay in review. Other tools here don't expose MCP interfaces.
What is “no token tax” in a feedback tool context?
Some tools price by events, API calls, or AI credits consumed, meaning busy projects or automated agents drive up costs unpredictably. Quell charges a flat per-user rate with no per-event metering: agents can read and write feedback as frequently as needed without the bill changing.
Is Userback free to use?
Userback has a genuine free tier limited to two seats with no time cap. It covers visual annotation and a shared inbox. REST API access requires upgrading to Business Plus at $23/seat/month (verified June 2026), a significant step up from the free or Team tier.
How does agent-native feedback differ from standard visual bug reporting?
Standard visual bug reporting is designed for humans: a reporter annotates a screenshot, a developer reads it and decides what to do. An agent-native feedback platform like Quell is designed so coding agents participate in that loop directly — they read structured context, make triage decisions, open issues, and comment, with a trust model that keeps humans in control of what gets auto-routed.