Cycling through all six. Tap any point to stop.
Measured on six things
Your knowledge, lit.
Security
52HTTPS, secure cookies, SPF and DMARC are in place, but HSTS, CSP, and frame protection are missing. No SOC 2 or bug bounty found. Table-stakes security for an early-stage tool.
Screenshots
1 / 1
About this app
Takibi Base (Takibi) is a shared knowledge base for AI agents. One human operator organizes documents into projects and folders and controls access with Profiles. Agents retrieve exact cited passages and an evidence support score via CLI or API; Takibi does not generate answers.
Key features
- 1Exact passage retrieval with citations and evidence support scores, no generated answers.
- 2Profile-based access control: choose which folders each agent key can read.
- 3CLI and API access with Node.js 22+ support, plus document conversion to Markdown.
Target customer profile
Drawn from the product itself, not from a survey.
Demographic
AI development teams
Pain points
Difficulty in providing reliable, citable knowledge to AI agents.
Primary needs
Centralized document management with precise citation capabilities.
Demographic
Customer support managers
Pain points
AI chatbots giving inaccurate or unsupported answers.
Primary needs
Access control to ensure agents only use approved documents.
Demographic
Legal and compliance officers
Pain points
Risk of AI agents accessing sensitive or unverified information.
Primary needs
Granular permissions and audit trails for AI agent access.
Scout Review
Written by AI from measured evidence, scored out of 100.
Takibi Base is a shared knowledge base for AI agents that returns exact cited passages instead of generated answers. The product is well-designed, fast, and documented with unusual care for an early-stage tool. Its core value proposition—giving agents access to approved documents with citations and an evidence support score—addresses a genuine pain point for teams deploying AI in support, legal, or compliance contexts. The pricing is straightforward: €18/month for 2 Profiles and 250 asks/day, or €52/month for 8 Profiles and 800 asks/day, both with a 7-day card-required trial. The main weaknesses are security headers (5/13 checklist points, no CSP or HSTS) and a narrow target market. Accessibility scores well on automated audits but lacks a public statement. Overall, it's a competent, focused tool that knows exactly what it is—and what it refuses to do.
Measure by measure
Usability
68Step-by-step docs and a sample workspace lower the barrier, but the 7-day trial requires a card upfront and the sample resets on refresh. Solid onboarding for a CLI-first tool.
Design
74Clean, modern dark UI with clear hierarchy, product mockups, and a polished pricing page. Docs use a command palette and structured guides. Above-average for an early-stage dev tool.
Speed
88Lighthouse mobile 99/100 (LCP 2.0s, CLS 0, TBT 0ms) and desktop 100/100. The site is fast and lightweight, with no reported lag complaints.
Security
52HTTPS, secure cookies, SPF and DMARC are in place, but HSTS, CSP, and frame protection are missing. No SOC 2 or bug bounty found. Table-stakes security for an early-stage tool.
Accessibility
80Lighthouse accessibility 100/100 with one failing audit on visible text labels. No a11y statement or multi-language support found. Strong automated score for an early-stage product.
Growth
58A real problem: AI agents need citable, access-controlled knowledge. The wedge is exact passages with citations and no answer generation, but the category is crowded and the target is narrow.
Improvement areas
- Security headers: add HSTS, CSP, frame protection, and other missing HTTP headers.
- Trial friction: remove the card requirement or offer a free tier with limited usage.
- Accessibility: publish a statement and fix the failing label audit for screen readers.
Design & Usability
Takibi Base presents a polished, modern dark interface with clear product mockups and a well-structured pricing page. The design language is consistent and professional, though it follows familiar SaaS patterns rather than breaking new ground. Usability is a mixed bag: the documentation is genuinely excellent, offering a step-by-step 'Your first retrieval' guide that walks through uploading a document, creating a Profile, connecting via CLI, and checking results. A sample workspace lets you explore without signing up, but uploads, imports, downloads, and billing all require a paid workspace, and the 7-day trial demands a credit card upfront. For a CLI-first tool aimed at developers, the onboarding is above average, but the card requirement adds friction before any value is demonstrated. The command palette in the docs is a nice touch that speeds up navigation.
Speed & Security
Speed is a standout strength. Lighthouse mobile performance scores 99/100 with an LCP of 2.0s, CLS of 0, TBT of 0ms, and FCP of 1.1s; desktop hits a perfect 100/100. The site is lightweight and snappy, with no user complaints about lag found in the collected evidence. Security is more modest. HTTPS is enforced, secure cookies are present, and SPF and DMARC records are configured, but HSTS, CSP, frame protection, X-Content-Type-Options, Referrer-Policy, and Permissions-Policy headers are all missing, yielding a 5/13 checklist score. No SOC 2, ISO certification, or bug bounty program was found. For a product that handles document access control and agent keys, the absence of a content security policy and clickjacking protection is a notable gap, though not unusual for an early-stage tool.
Accessibility & Growth
Accessibility scores 100/100 on Lighthouse's automated audit, with one failing check: elements with visible text labels do not have matching accessible names. No accessibility statement or multi-language support was found on the site. The automated score is strong, but manual testing would be needed to confirm real-world screen reader compatibility. Growth potential is moderate. Takibi Base targets a real problem: AI agents need reliable, citable knowledge with granular access control. Its differentiator is that it returns exact passages with citations and an evidence support score rather than generating answers, which appeals to teams worried about hallucinations. However, the category is crowded with vector databases like Pinecone and Weaviate, and the target segment (AI development teams, support managers, legal/compliance) is narrow. The domain was registered in September 2026, and social profiles are limited to Twitter and YouTube, indicating an early-stage product with room to grow but no clear network effects yet.
Conclusion
If your AI agents are confidently making things up and you need them to cite their sources, Takibi Base is worth a look. The CLI-first approach and exact-passage retrieval are a good fit for development teams already comfortable with terminal workflows. The 7-day trial with a card requirement is a hurdle, but the sample workspace and thorough docs let you evaluate the core retrieval experience before committing. Just don't expect it to write your answers—that's your agent's job. For teams that need auditable, access-controlled knowledge retrieval, this is a solid early-stage option. For everyone else, the vector database incumbents may be enough.
Suggestions for improvement
- Add HSTS, CSP, and frame protection headers to harden the web surface.
- Offer a no-card trial or a free tier with limited asks to reduce signup friction.
- Publish an accessibility statement and fix the failing label audit.
- Expand integrations beyond CLI/API with pre-built connectors for popular agent frameworks.
Alternatives comparison
Named competitors, point by point. Nobody paid to appear here or to be left out.
| Primary function | Shared knowledge base returning exact cited passages for AI agents | Vector database for similarity search | Open-source vector search engine |
|---|---|---|---|
| Built-in citations | Yes, returns document and chunk citations with every passage | No, requires custom implementation | No, requires custom development |
| Access control for agents | Profiles with folder-level permissions and key management | Namespace and API key based, not document-level | Requires custom authorization layer |
| Answer generation | No, returns only source passages | No, retrieval only | No, retrieval only |
| Pricing model | €18/month (2 Profiles, 250 asks/day) or €52/month (8 Profiles, 800 asks/day) | Free tier; paid plans from $25/month | Free open-source; cloud from $25/month |
Pinecone
Vector database for AI applications, offering similarity search but lacking built-in citation and access control features.
Weaviate
Open-source vector search engine with modular architecture, suitable for AI knowledge retrieval but requires custom development for citations.
Comparing options? See Takibi Base alternatives, scored side by side
More Productivity apps
View all Productivity
Qamarero
92An all-in-one hospitality software for restaurants and ba...

Puzzly – B2B Lead Generation & Sales Intelligence Platform
92Puzzly is an all-in-one B2B lead generation platform that...

The Stack Newsletter
92A weekly newsletter delivering curated tech jobs, intervi...

StoryPad
92Privacy-first open source journal app with a continuous t...

Vibe Coding Agent Skills
92A free, open-source platform offering 110 AI agent skills...

Notion
92All-in-one AI workspace for docs, projects, and knowledge...

EmbedMyReviews
92White-label reputation management software for agencies s...

QRBliss
91Branded QR codes with AI Brand Sync, 15 free dynamic code...

Strukchad Property Management
91Property management for Kenyan letting agents: rent invoices raised every month, M-Pesa rent prompts to tenants, arrears tracking and landlord statements.

CC Switch
91Open-source desktop app for switching AI coding providers...

Gerty
91Private, local AI assistant for Mac that runs offline

BACKLNK
90Backlnk is a free SEO and AI visibility website: 15 website checks, a full audit as a PDF report, and guides to get found in search and AI answers.
A platform dedicated to providing unbiased reviews of newly launched applications, analyzing everything from their features to their full potential.
info@scoutforge.net© 2026 Scoutforge. All rights reserved.