ProofHound
ProofHound automates prompt optimization for LLM classification workflows.
What ProofHound is
ProofHound automates prompt optimization for LLM classification workflows. It analyzes error cases, iterates prompts, validates results, and supports deployment and rollback. Teams can upload CSV, TSV, JSONL, JSON array, or ZIP datasets, then map fields in the interface. Optimization can target overall accuracy or category-specific goals, such as recall for high-risk categories or precision for error-prone classes. Immutable prompt versions preserve variable settings, output rules, and differences between versions. Experiment records include datasets, model configurations, sample judgments, and overall and category metrics. Deployment options include gray traffic release, A/B testing, full rollout, and one-click rollback. ProofHound connects through its web UI, webhooks, API tokens, and MCP. Users bring their own model provider; ProofHound says it does not charge per model call. A free plan is listed as free, with limits of three projects, one member, three concurrent LLM calls, 5GB retained storage, and 200MB per dataset upload. A self-hosted open-source option is also listed as free.
Good
- Accepts CSV, TSV, JSONL, JSON array, and ZIP datasets.
- Supports accuracy and category-specific optimization targets.
- Records immutable prompt versions and experiment details.
- Offers gray releases, A/B testing, and rollback.
- Self-hosted edition supports private data storage.
Watch out
- Free plan limits projects to three and members to one.
- Free plan allows three concurrent LLM calls.
- Free dataset uploads are limited to 200MB.
- Pro checkout is not open yet.
Who it suits
ProofHound suits operations, risk, and analyst teams working on classification tasks such as risk control, financial judgment, content moderation, or customer service intent recognition. The self-hosted edition may suit teams that need private deployment and data storage on their own infrastructure.
Verdict
ProofHound combines prompt tuning, experiment records, and deployment controls for classification workflows. Its free and self-hosted options are listed as free, while the planned Pro plan is not yet open for checkout.
Plans and prices
From ProofHound's own pricing page, checked 2 Oct 2026.
| Plan | Price | What it gives you |
|---|---|---|
| Free | Free | 3 projects · 1 member · 3 concurrent LLM calls · 5GB retained project storage · 200MB per dataset upload · $0 CNY 0 · per organization |
| Self-hosted open source | Free | Full core capabilities · own infrastructure · custom model integration · private data storage · single workspace deployment · community support · Free forever |
| Pro | $29 a month | Unlimited projects and members under shared org quota · 50 concurrent LLM calls · 7-day workflow runtime · 50GB retained project storage · 2GB per dataset upload · full RBAC and integration channels · checkout not open yet · Coming soon · planned $29/mo |
The facts
Each one read from ProofHound's own pages, with the date it was checked.
What it does
- Dataset inputs
- The product supports CSV, TSV, JSONL, JSON array, and ZIP dataset uploads, with flexible field mapping in the UI.checked 2 Oct 2026
- Experiment audit
- The platform records prompt versions, datasets, model configurations, sample judgments, and overall and category metrics for traceable and reproducible experiments.checked 2 Oct 2026
- Integrations
- The site lists Web UI, Webhook, API Token, and MCP as connection options for business systems and AI agents.checked 2 Oct 2026
- Intended users
- The product targets critical classification workflows including risk control, financial judgment, content moderation, and customer service intent recognition, including low-code work by operations, risk, and analyst teams.checked 2 Oct 2026
- Model usage
- Customers bring their own model provider, and ProofHound says it does not charge per model call.checked 2 Oct 2026
- Optimization targets
- Users can optimize overall accuracy or tune category-specific metrics such as recall for high-risk categories and precision for error-prone classes.checked 2 Oct 2026
- Production controls
- ProofHound supports gray traffic release, A/B testing, full rollout, and one-click rollback for prompt deployments.checked 2 Oct 2026
- Purpose
- ProofHound automates prompt optimization for LLM classification tasks by analyzing error cases, iterating prompts, validating results, and supporting deployment and rollback.checked 2 Oct 2026
- Roadmap
- Evaluation, comparison, and optimization for generative LLM tasks and ProofHound Cloud Managed Enterprise Edition are listed as upcoming.checked 2 Oct 2026
- Security and data control
- The self-hosted edition supports private deployment and private data storage on the user's own infrastructure; hosted data is retained until user deletion subject to plan quota.checked 2 Oct 2026
- Support
- The site lists GitHub, Discord, a Chinese-speaking QQ group, and email for community discussion, product updates, and business contact.checked 2 Oct 2026
- Version history
- ProofHound records immutable prompt versions with variable configurations, output rules, and version differences.checked 2 Oct 2026
Alternatives to ProofHound
- 1
Langfuse
Free planOpen sourceWebAPILinuxSelf-hosted8.2From $29 a month - 2
LangSmith
Free planWebAPILinuxmacOS8.0$39 usage-based - 3
SpacePrompts
Free planWebAPIExtension7.4From $5 a month - 4
PromptLayer
Free planWebAPISelf-hosted7.0From $49 a month - 5
PromptHub
Free planWebAPI6.8From $12 a month - 6
Promptitude
Free planWebAPI6.7From $39 a month