Free
Free
Open-source library with free pip and conda installs documented on the homepage
- Access
- Open-source library with free pip and conda installs documented on the homepage
Finding the possibilities
captum.ai
● Identity checkedAssistant · Vendor site 2026-10-04
Captum is Meta’s open-source PyTorch library for model interpretability, offering algorithms such as Integrated Gradients, DeepLift, and Layer Conductance. Researchers install it via pip or conda to attribute predictions across vision, text, and multimodal models with minimal network changes.
- Category
- Assistant
- Free access
- Open-source library with free pip and conda installs documented on the homepage
- API access
- Not confirmed
Is captum.ai right for you?
A good fit for
ML engineers and researchers who need extensible interpretability methods inside PyTorch training pipelines.
Before you choose
Captum is a developer library—not a hosted SaaS—so support comes via docs, tutorials, and GitHub rather than SLAs.
Model interpretability for PyTorch
Captum provides an extensible open-source toolkit for understanding neural network predictions with tutorials and API reference docs.
What it can do
Features & capabilities
Unknown is different from unavailable. Each fact carries its own evidence.
| Capability | Value | Evidence | Checked |
|---|---|---|---|
| PyTorch native | Homepage states Captum works with most PyTorch models with minimal modifications. | Facts sourced | 2026-10-04 |
| Multimodal | Feature list covers interpretability for vision, text, and other modalities. | Facts sourced | 2026-10-04 |
| Install | Docs link shows conda and pip install commands on the landing page. | Facts sourced | 2026-10-04 |
Understand the total cost
captum.ai pricing & plans
Alternatives to captum.ai
View all ↗
InthInth embeds cookie consent and privacy scanning directly in your application. It offers free scans that show issue counts before purchase, continuous PR and runtime monitoring with Inth Agent, and consent infrastructure with regional policy packs and session limits.
Explore tool
pingpong.onePingPong is a status page platform for publishing component health, incidents, and subscriber updates with custom domains and integrations to chat and on-call tools. Pricing offers a free tier plus Solo, Pro, and Business plans with increasing metrics, team seats, and subscribers.
Explore tool
Carrot LabsSuperPenguin, from Carrot Labs, is LLM cost management software that attributes AI spend by customer, feature, and team while unifying bills from OpenAI, Anthropic, Gemini, Cursor, and other providers. It adds Coding ROI views for agent spend per merged pull request and a Mac menu-bar app for live agent usage.
Explore toolThe practical questions
Frequently asked questions
What is captum.ai used for?
Captum is Meta’s open-source PyTorch library for model interpretability, offering algorithms such as Integrated Gradients, DeepLift, and Layer Conductance. Researchers install it via pip or conda to attribute predictions across vision, text, and multimodal models with minimal network changes.
Does captum.ai have a free plan?
Open-source library with free pip and conda installs documented on the homepage. This record lists ongoing free access; check the plan limits before starting.
How much does captum.ai cost?
No paid monthly price is listed; this record treats the product as free to start. See the plan cards for entitlements, billing commitments and seat minimums.
Can I use captum.ai through an API?
Not confirmed. API access and subscription access may have different terms; consult the linked sources.
What should I check before choosing it?
Captum is a developer library—not a hosted SaaS—so support comes via docs, tutorials, and GitHub rather than SLAs.
Price history
No retained pricing changes yet. A current price alone does not establish a historical trend.
How this profile is supported
Identity and profile
Reviewed vendor source
Read original source ↗Facts apply to the named version and check date. Send a sourced correction if something changed.