M/s Ashoka Infra
₹2.31 Cr
84.1
For faster, secure public procurementReduce the tender
Reduce the tender
lifecycle from
months to hours.
Scribe drafts the tender, runs the CVC pre-check, and tracks every approval.
Aligned with India's procurement framework
GFR 2017CVC GuidelinesMeitY Empanelled CloudGIGW 3.0 / UX4G
How it works
Six stages. One system.
01 · DPR Creation
The BOQ builds itself.
The Bill of Quantities is generated automatically, using AI-suggested items with a defensible, market-rate cost basis. Cost mismatches are flagged before submission.
02 · Tender Creation
The tender writes itself, section by section.
The tender is drafted section by section from a plain description of the work.
03 · Pre-bid Queries
Hundreds of queries. Grouped by AI.
Similar questions are grouped into themes automatically, with one consistent, officer-approved answer drafted per group.
04 · Bid Evaluation
Every score, backed by evidence.
PQ and TQ criteria are checked against submitted documents.
05 · Bid Comparison
The comparison table builds itself, live.
As each bid moves through evaluation, a live comparative table updates in real time. The full ranking is ready the moment evaluation ends.
06 · Execution
From award to a signed, closed file.
Milestones, running-account bills, and validations are tracked through delivery.
What Scribe changes
Under 1 minute
to draft a complete tender, section by section
30min
to process 200+ pre-bid queries, reduced from 3–4 days
100%
audit trail on every action and decision
Meet Scribe
One platform. The complete procurement lifecycle.
From tender drafting to Bidder Compliance, every step is AI-assisted, officer-approved, and audit-ready.
Tender Creation
Describe the work. Scribe writes the tender.
CVC risk check
AI flags what a reviewer will raise.
Pre-bid Queries
212 queries. 5 themes. 30 minutes.
Bid Evaluation
Every mark, cited.
Comparative Statement
L1 in one look.
Guided wizard to assist at each step.
AI picks up the input details and drafts the tender.
AI drafts, officer decidesEvery suggestion is advisory.
CVC pre-check on every tenderPredicts what a reviewer is likely to flag.
Available in English, Hindi, or MarathiGenerated with one click.
Rule-based, not guesswork
Compliance you can defend
The CVC pre-check is deterministic and rule-based, not LLM-generated. It follows GFR 2017, CVC guidelines, and the Manual for Procurement.
GFR 2017 · Rule 170CVC pre-checkManual for Procurement
Self-hosted LLM
Your data never leaves India
Scribe runs a self-hosted model on MeitY-empanelled cloud, entirely within India. No third-party AI API ever sees a single bid.
MeitY-empanelled cloudDPDP ActIT Act 2000
Domain-trained
Trained on 1,000+ tenders
The model learned from over a thousand real government tenders, so its drafts and checks speak the language of Indian public procurement.
DSR-awareGFR-awareSector templates
Nothing off the record
Every access, logged
Role-based access, quotes masked until the formal reveal, and an immutable audit trail of who did what, and when.
Role-based accessImmutable audit trailMasked quotes
Straight answers, kept simple.
Scribe touches confidential bids, public money, and audits, so these questions deserve direct answers. The following are the most common; further detail is available in a live walkthrough.
Reach outDoes Scribe replace the officer or the approval process?
No. Scribe functions as a co-pilot: it drafts, checks, and tracks, while every compliance pointer remains advisory and all signings and approvals happen offline. The decision and accountability stay with the officer and the competent authority.
How does it keep confidential bid data safe?
Access is role-based: each role sees only what it is permitted to view. Quote amounts remain masked until formally revealed, in a deliberate two-step action. A Super Admin audit view logs who accessed confidential data, and when.
Which compliance norms does Scribe align with?
The workflow follows GFR 2017, CVC guidelines, and the Manual for Procurement of Works, Goods, and Services. The consolidated CVC pre-check anticipates issues a reviewer is likely to raise. Compliance checks are rule-based rather than LLM-generated, and deterministic where accuracy is critical.
What happens to department data? Does the AI use it?
Scribe runs a self-hosted Large Language Model (LLM) on MeitY-empanelled cloud infrastructure entirely within India. No third-party AI API has access to this data at any point. The platform is designed to comply with the DPDP Act and IT Act 2000, with data sovereignty built in from the outset.
Can it generate documents in Hindi and Marathi?
Yes. Tenders can be generated in English, Hindi, or Marathi, with controls for logo and seal placement and support for adding custom sections specific to departmental requirements.
Is the portal accessible for all officers, including those with disabilities?
Scribe is designed to meet GIGW 3.0, UX4G guidance, and WCAG 2.1 AA, including full keyboard operation, visible focus indicators, sufficient colour contrast (AA+ on all text), and screen-reader compatibility.