Federal Contracts API for Developers

SAM.gov API Response Structure Reference

Technical reference for SAM.gov's JSON response format. This guide covers nested fields, data types, parsing patterns, and helper code for working with federal contract data.

About this guide: SAM.gov returns detailed nested JSON. This reference helps you understand the structure and write effective parsing code.
Missing Contact Information?

SAM.gov provides opportunity data. Get verified decision maker contacts (email, phone, LinkedIn) to reach the right people. Try GovCon Contacts →

The SAM.gov JSON Response Structure

Example: Single Opportunity Response (Heavily Nested)

Here's a simplified version of what SAM.gov returns for a single contract opportunity:

{"opportunitiesData": [ {"noticeId":"abc123def456ghi789","title":"Software Development Services","sol":"W52P1J-25-R-0001","fullParentPathName":"Department of Defense.Department of the Army.Army Contracting Command.Army Contracting Command - Detroit Arsenal (ACC-DTA)","fullParentPathCode":"DOD.DA.ACC.ACC-DTA","postedDate":"2025-11-01","type":"Solicitation","baseType":"Solicitation","archiveType":"auto15","archiveDate":"2025-12-16","typeOfSetAsideDescription":"Total Small Business Set-Aside (FAR 19.5)","typeOfSetAside":"SBA","responseDeadLine":"2025-11-30T17:00:00-05:00","pointOfContact": [ {"type":"primary","title":"","fullName":"John Smith","email":"[email protected]","phone":"586-555-1234","fax": null }, {"type":"secondary","title":"Contracting Officer","fullName":"Jane Doe","email":"","phone":"586-555-5678","fax": null } ],"placeOfPerformance": {"streetAddress":"123 Main Street","streetAddress2":"Suite 100","city": {"code":"48397","name":"Warren" },"state": {"code":"MI","name":"Michigan" },"zip":"48397","country": {"code":"USA","name":"UNITED STATES" } },"organizationType":"OFFICE", // String, not object in current API"naicsCode":"541511", // String format"naicsCodes": ["541511"], // Also available as array"additionalInfoLink":"https://sam.gov/opp/abc123def456ghi789/view","uiLink":"https://sam.gov/opp/abc123def456ghi789/view","links": [ {"rel":"self","href":"https://api.sam.gov/prod/opportunities/v2/search?noticeid=abc123def456ghi789&limit=1" } ],"resourceLinks": [ {"type":"document","name":"Amendment 001","link":"https://sam.gov/api/prod/opps/v3/opportunities/resources/files/abc123/download?&token=..." }, {"type":"document","name":"Original Solicitation","link":"https://sam.gov/api/prod/opps/v3/opportunities/resources/files/def456/download?&token=..." } ],"officeAddress": {"zipcode":"48397","city":"Warren","countryCode":"USA","state":"MI" }, // Award information (if available) - completely separate structure"award": {"date":"2025-12-15","number":"W52P1J-25-C-0001","amount": 150000,"lineItemNumber":"0001","awardee": {"name":"ACME Software Solutions","location": {"streetAddress":"456 Tech Drive","city":"Detroit","state":"MI","zipCode":"48201","countryCode":"USA" },"ueiSAM":"ABC123DEF456","cageCode":"1A2B3" } } } ],"totalRecords": 1247,"offset": 0,"limit": 10 }

What the data actually looks like at scale

The nested structure above is only half the story. Once you have a few hundred thousand notices in a database, you start finding problems the schema doesn't warn you about. These are real defects we have observed across SAM contract data, not theoretical edge cases.

Snapshot as of May 13, 2026. Our index currently holds 255,189 opportunity notices (including 56,590 award notices), 163,471 exclusion records, and 872,793 SAM entity registrations. Counts grow as SAM publishes new notices; the structural ratios below (UEI coverage gap, amendment churn, etc.) are stable across months.

Implausible award dates

Among 56,590 award notices in our index, 41 records have award_date values past 2027, including dates in years 3025, 6202, 7202, and 9202. These are typos that survived every upstream check; one record is dated nearly seven thousand years in the future. Less than 0.1% of records, but enough to break a "top N by date" analytics query in production.

What we do: validate every date param at the API edge before it reaches the database. Customers got a 500 when they accidentally sent 2026-06-31 (June has 30 days) until we patched our own validator; we now reject month-day combinations that do not exist on the calendar, with a clear 400 error.

Missing buyer-identification on award notices

14.4% of award notices have no award_uei_sam populated. 8,171 of 56,590 records cannot be linked back to a vendor by UEI. Some are pre-UEI legacy records; some are GSA Schedule and IDV vehicles where the awardee is captured only as a freetext name string. If you build a "company profile" feature and join on award_uei_sam, you will silently undercount by roughly one in seven.

Published-already-archived notices

SAM occasionally publishes contract opportunities whose archiveDate is on or before the postedDate. In the last 52 days alone we have ingested 399 such records. They are technically in the dataset but never visible in any live open opportunities feed, because the archive cutoff has already passed at the moment of publication. If you depend on SAM's "active" semantics rather than computing your own from archive_date_detailed, you will miss these notices entirely.

Notice IDs are not stable across amendments

noticeId looks like a primary key but does not behave like one. SAM mints a new noticeId for every amendment to a solicitation. Our internal audits show roughly 74% of records that appear "missing" between two snapshots of the dataset are actually older revisions of solicitations already present under a newer noticeId. If you treat the field as a stable identifier and dedupe on it, you will accumulate every revision as a separate record.

The more stable identifier is solicitation_number (the human-readable string like W52P1J-25-R-0001), but it is not unique across agencies. Real deduplication needs both fields plus a posted-date window.

solicitation_number and notice_id get confused constantly

The two fields use entirely different formats. noticeId is a 32-character hex UUID (abc123def456ghi789...). solicitation_number follows agency-specific patterns like N0003925R4014 or 36C25026B0043. Customers regularly paste the wrong one into URL paths and get back HTTP 400 or empty results. We see this pattern routinely in our access logs. Both fields belong in any UI you build on this data, clearly labeled.

The literal string "null" instead of a SQL NULL

SAM's exclusion records ship with the four-character string "null" in fields that should be empty. At an earlier point we found that 119,343 of 163,450 exclusion rows in our staging database had uei_sam stored as the literal text "null" rather than SQL NULL. A literal lookup at /exclusions/null matched a real record; a search filter ?uei_sam=null matched 73% of the table. We have since coerced "null", "none", "n/a", and the empty string to actual SQL NULL during ingest, and our API no longer exhibits this behavior. Anyone loading the same source data into their own database will need to do the same normalization, or their WHERE column IS NULL queries will silently miss everything.

Inconsistent date formats across fields

Within a single opportunity record, SAM mixes formats:

The CSV exports use yet another format. A robust parser has to recognize at least four patterns before sorting and filtering work correctly.

Mixed types in the same field

naicsCode can be a string (single primary code) or an array, depending on the record. placeOfPerformance.city can be an object with {code, name} or a plain string. set_aside_type may be the code ("SBA") or the description ("Total Small Business Set-Aside (FAR 19.5)") depending on how the source system reported it. Your parser needs isinstance() checks at every level.

Free-text contamination in structured fields

awardee_name sometimes contains the full mailing address concatenated into the name itself ("ACME CORP 123 MAIN ST CHESAPEAKE VA 23320-5999"). Splitting this back into name + address with a regex butchers 21% of records on the messy variants. The safest move is to keep the raw string and accept that "company name" is sometimes "company name plus a comma-separated address." Searching by name needs full-text or trigram indexing rather than equality.

Cancellation status is not exposed

SAM has three lifecycle states for a notice: active, archived, and cancelled. The JSON exposes only the first two. A cancelled notice still has a future archiveDate and looks active in the data; the cancellation lives in a separate event stream or shows up only when SAM updates the human-readable detail page. If you need cancellation awareness, you have to scrape it.

Summary of findings (May 13, 2026 snapshot)

Issue Scale we observe Customer impact
Implausible award_date values 41 of 56,590 award notices Breaks "sort by date" analytics
Missing award_uei_sam 14.4% (8,171 of 56,590) Vendor-profile join misses roughly 1 in 7
Published-already-archived notices 399 in the last 52 days Invisible in any "open opportunities" feed
Notice ID churn on amendments ~74% of apparent "missing" records Duplicate records if deduped on the wrong field
Literal "null" strings (historical, since cleaned in our index) Was 119,343 of 163,450 exclusion uei_sam rows WHERE uei_sam IS NULL silently misses everything
Mixed date formats in one record Every opportunity Sorting and filtering require multi-format parsing
Mixed types per field Common Runtime type errors
Address concatenated into awardee_name Roughly 1 in 5 of the messier variants Regex splitting corrupts vendor names
No cancellation status in the JSON All notices Cancelled notices look active until archived

None of these are documented anywhere in SAM's official references. We surface them here because they cost real engineering time to discover on your own dataset.

The public bulk is a filtered subset of SAM's full registry

The monthly Public V2 entity extract that most tools (including ours) build on contains 876,399 entities as of the May 2026 release. SAM's internal search index, accessible via sgs/v1/search?index=ent, returns 2,332,884 records for the same query. The 1.46M-record gap isn't a bug or a stale snapshot. The bulk is a deliberately filtered subset.

Filter on sgs/v1/search?index=entRecordsNotes
(no filter, full registry)2,332,884What SAM indexes internally
publicDisplayFlag = "Y"1,333,263Opt-in public disclosure only
recordStatus = "A"848,808Close to the 782K active rows in the public bulk
Public V2 monthly bulk (all filters applied)876,399What ends up in our table

Where the extra 1.46M comes from

Most of the non-bulk-eligible records cluster into three buckets.

Privacy-redacted entities (roughly 1.0M). SAM lets registrants opt out of public disclosure at registration time, setting publicDisplayFlag="N". They stay in SAM's index so the website can show "this UEI is registered, just not browsable", but no public API or bulk extract surfaces them. They are FOIA-protected by design.

Non-contract-vendor registration purposes. The bulk is filtered to entities registered for All Awards (full federal contracting). SAM also indexes registrations purposed only for Federal Assistance Awards (grants), IRS 1099 reps and certs, or sub-vendor representations. These exist in ent but are not included in the contract-vendor bulk because they are not bidding on contracts.

Pre-UEI / CCR-era migration debris. SAM replaced CCR (Central Contractor Registration) in 2018, and DUNS+4 was replaced by UEI in 2022. Old CCR records were migrated into SAM's index for audit-trail continuity. About 77% of records in sgs/v1/search?index=ent still carry the 2012-07-28 migration-baseline timestamp on the internal emrValue field, meaning they haven't been touched since the CCR to SAM migration. Mostly defunct entities retained for reference.

What this means in practice

If you are querying federal contractors for vendor research, due diligence, or contract pipeline intelligence, the 876K public-bulk universe is the right scope. The extra 1.46M in SAM's full index is noise for that use case: roughly 1.0M privacy-redacted (legally inaccessible to any API), about 400K migration debris (defunct companies, decade-old inactive records), and 50K to 100K non-contract-vendor registrations.

The only meaningful gap is the privacy-redacted cohort: roughly 1M actively-contracting entities who chose not to be publicly listed. No public API surfaces them. Awareness of this gap matters when reconciling totals against counts from other sources (USASpending's vendor count derives from a different pipeline and can differ in turn).

Key Parsing Considerations

Common Patterns to Handle

  1. Nested Objects: Data can be 3-4 levels deep (e.g., placeOfPerformance.city.name)
  2. Optional Fields: Always use null-safe access patterns
  3. Mixed Types: Some fields may be string, array, or object depending on data
  4. Contact Arrays: pointOfContact contains multiple contacts with different types
  5. NAICS Codes: Returns BOTH naicsCode (string) AND naicsCodes (array)
  6. Address Formats: placeOfPerformance vs officeAddress use different structures
  7. Date Formats: Mix of date-only (YYYY-MM-DD) and ISO timestamps
  8. Set-Aside Codes: Use lookup table to translate codes to readable names
  9. Agency Hierarchy: fullParentPathName contains dot-separated department chain

Real Parsing Code: 400+ Lines Required

Production Parsing Function (Partial Example)

Here's what you actually need to write to parse SAM.gov responses reliably:

import re from datetime import datetime, timezone from typing import Dict, List, Optional, Any class SAMOpportunityParser:"""Complex parser for SAM.gov opportunity data""" def __init__(self): # Set-aside code translations self.set_aside_codes = { 'SBA': 'Small Business Set-Aside', 'A6': '8(a) Set-Aside', 'HZC': 'HUBZone Set-Aside', 'SDVOSBC': 'Service-Disabled Veteran-Owned Small Business', 'WOSB': 'Women-Owned Small Business', 'EDWOSB': 'Economically Disadvantaged Women-Owned Small Business', '': 'Full and Open Competition' } # Organization type mappings self.org_type_codes = { 'O': 'Office', 'D': 'Department', 'A': 'Agency', 'S': 'Sub-Agency' } def parse_opportunity(self, raw_data: Dict) -> Dict:"""Parse a single opportunity from SAM.gov response""" try: # Extract basic fields with null checking opportunity = { 'notice_id': self._safe_get(raw_data, 'noticeId'), 'title': self._safe_get(raw_data, 'title', '').strip(), 'solicitation_number': self._safe_get(raw_data, 'sol'), 'posted_date': self._parse_date(raw_data.get('postedDate')), 'response_deadline': self._parse_datetime(raw_data.get('responseDeadLine')), 'notice_type': self._safe_get(raw_data, 'type'), 'base_type': self._safe_get(raw_data, 'baseType'), 'archive_date': self._parse_date(raw_data.get('archiveDate')), } # Parse complex agency hierarchy agency_data = self._parse_agency_hierarchy(raw_data) opportunity.update(agency_data) # Parse set-aside information opportunity['set_aside'] = self._parse_set_aside(raw_data) # Parse contact information (complex nested array) opportunity['contacts'] = self._parse_contacts(raw_data.get('pointOfContact', [])) # Parse performance location (deeply nested) opportunity['performance_location'] = self._parse_location( raw_data.get('placeOfPerformance', {}) ) # Parse office address (different structure than performance location) opportunity['office_address'] = self._parse_office_address( raw_data.get('officeAddress', {}) ) # Parse NAICS codes (array of objects) opportunity['naics_codes'] = self._parse_naics_codes( raw_data.get('naicsCode', []) ) # Parse organization type opportunity['organization_type'] = self._parse_organization_type(raw_data) # Parse resource links (documents, amendments) opportunity['resource_links'] = self._parse_resource_links( raw_data.get('resourceLinks', []) ) # Parse award information (if available) opportunity['award_info'] = self._parse_award_info( raw_data.get('award', {}) ) # Generate clean URLs opportunity['sam_url'] = self._generate_sam_url(opportunity['notice_id']) # Extract additional metadata opportunity['total_records'] = raw_data.get('totalRecords') return opportunity except Exception as e: # Robust error handling for malformed data print(f"Error parsing opportunity {raw_data.get('noticeId', 'unknown')}: {e}") return self._create_error_record(raw_data, str(e)) def _safe_get(self, data: Dict, key: str, default: Any = None) -> Any:"""Safely extract value with null checking""" value = data.get(key, default) if isinstance(value, str): return value.strip() if value else default return value if value is not None else default def _parse_agency_hierarchy(self, data: Dict) -> Dict:"""Parse complex agency hierarchy string""" full_path = data.get('fullParentPathName', '') path_code = data.get('fullParentPathCode', '') # Split hierarchy:"Dept.Agency.Sub-Agency.Office" path_parts = full_path.split('.') code_parts = path_code.split('.') return { 'department': path_parts[0] if len(path_parts) > 0 else '', 'agency': path_parts[1] if len(path_parts) > 1 else '', 'sub_agency': path_parts[2] if len(path_parts) > 2 else '', 'office': path_parts[3] if len(path_parts) > 3 else '', 'department_code': code_parts[0] if len(code_parts) > 0 else '', 'agency_code': code_parts[1] if len(code_parts) > 1 else '', 'full_agency_name': full_path, 'full_agency_code': path_code } def _parse_set_aside(self, data: Dict) -> Dict:"""Parse set-aside information with code translation""" code = data.get('typeOfSetAside', '') description = data.get('typeOfSetAsideDescription', '') return { 'code': code, 'description': description, 'standardized_name': self.set_aside_codes.get(code, code), 'is_small_business': code in ['SBA', 'A6', 'HZC', 'SDVOSBC', 'WOSB', 'EDWOSB'] } def _parse_contacts(self, contacts_data: List[Dict]) -> List[Dict]:"""Parse contact array with inconsistent structure""" contacts = [] for contact in contacts_data: if not isinstance(contact, dict): continue parsed_contact = { 'type': contact.get('type', '').lower(), 'title': self._safe_get(contact, 'title', ''), 'name': self._safe_get(contact, 'fullName', ''), 'email': self._clean_email(contact.get('email', '')), 'phone': self._clean_phone(contact.get('phone', '')), 'fax': self._clean_phone(contact.get('fax', '')) } # Skip contacts with no useful information if parsed_contact['name'] or parsed_contact['email']: contacts.append(parsed_contact) return contacts def _parse_location(self, location_data: Dict) -> Dict:"""Parse complex nested location structure""" if not location_data: return {} # Handle nested city/state/country objects city_obj = location_data.get('city', {}) state_obj = location_data.get('state', {}) country_obj = location_data.get('country', {}) return { 'street_address': self._safe_get(location_data, 'streetAddress', ''), 'street_address_2': self._safe_get(location_data, 'streetAddress2', ''), 'city': city_obj.get('name', '') if isinstance(city_obj, dict) else str(city_obj), 'city_code': city_obj.get('code', '') if isinstance(city_obj, dict) else '', 'state': state_obj.get('code', '') if isinstance(state_obj, dict) else str(state_obj), 'state_name': state_obj.get('name', '') if isinstance(state_obj, dict) else '', 'zip_code': self._safe_get(location_data, 'zip', ''), 'country': country_obj.get('code', '') if isinstance(country_obj, dict) else str(country_obj), 'country_name': country_obj.get('name', '') if isinstance(country_obj, dict) else '' } def _parse_office_address(self, office_data: Dict) -> Dict:"""Parse office address (different structure than performance location)""" if not office_data: return {} return { 'city': self._safe_get(office_data, 'city', ''), 'state': self._safe_get(office_data, 'state', ''), 'zip_code': self._safe_get(office_data, 'zipcode', ''), 'country': self._safe_get(office_data, 'countryCode', '') } def _parse_naics_codes(self, naics_data) -> List[Dict]:"""Parse NAICS - SAM.gov returns BOTH naicsCode (string) AND naicsCodes (array)""" naics_codes = [] # SAM.gov returns naicsCode as a string (e.g.,"541511") # AND naicsCodes as an array (e.g., ["541511"]) if isinstance(naics_data, str) and naics_data: naics_codes.append({ 'code': naics_data, 'title': '', # Title not provided by SAM.gov API 'is_primary': True }) elif isinstance(naics_data, list): # Handle legacy array format if encountered for naics in naics_data: if isinstance(naics, dict): parsed_naics = { 'code': self._safe_get(naics, 'code', ''), 'title': self._safe_get(naics, 'title', ''), 'is_primary': len(naics_codes) == 0 } if parsed_naics['code']: naics_codes.append(parsed_naics) elif isinstance(naics, str): naics_codes.append({ 'code': naics, 'title': '', 'is_primary': len(naics_codes) == 0 }) return naics_codes def _parse_award_info(self, award_data: Dict) -> Optional[Dict]:"""Parse award information (completely different structure)""" if not award_data: return None # Parse awardee information (nested in award object) awardee_data = award_data.get('awardee', {}) awardee_location = awardee_data.get('location', {}) return { 'award_date': self._parse_date(award_data.get('date')), 'award_number': self._safe_get(award_data, 'number', ''), 'award_amount': self._parse_amount(award_data.get('amount')), 'line_item_number': self._safe_get(award_data, 'lineItemNumber', ''), 'awardee_name': self._safe_get(awardee_data, 'name', ''), 'awardee_uei': self._safe_get(awardee_data, 'ueiSAM', ''), 'awardee_cage_code': self._safe_get(awardee_data, 'cageCode', ''), 'awardee_address': { 'street': self._safe_get(awardee_location, 'streetAddress', ''), 'city': self._safe_get(awardee_location, 'city', ''), 'state': self._safe_get(awardee_location, 'state', ''), 'zip_code': self._safe_get(awardee_location, 'zipCode', ''), 'country': self._safe_get(awardee_location, 'countryCode', '') } } def _parse_date(self, date_str: Optional[str]) -> Optional[str]:"""Parse various date formats from SAM.gov""" if not date_str: return None try: # Handle multiple date formats for fmt in ['%Y-%m-%d', '%m/%d/%Y', '%Y-%m-%dT%H:%M:%S']: try: dt = datetime.strptime(date_str.split('T')[0], fmt) return dt.strftime('%Y-%m-%d') except ValueError: continue return date_str # Return original if parsing fails except Exception: return None def _parse_datetime(self, datetime_str: Optional[str]) -> Optional[str]:"""Parse datetime with timezone handling""" if not datetime_str: return None try: # Remove timezone suffix for parsing clean_dt = re.sub(r'[-+]\d{2}:\d{2}$', '', datetime_str) dt = datetime.fromisoformat(clean_dt) return dt.isoformat() except Exception: return datetime_str def _clean_email(self, email: str) -> str:"""Clean and validate email addresses""" if not email: return '' email = email.strip().lower() # Basic email validation if '@' in email and '.' in email.split('@')[-1]: return email else: return '' def _clean_phone(self, phone: str) -> str:"""Clean phone number format""" if not phone: return '' # Remove non-numeric characters except + clean_phone = re.sub(r'[^\d+\-\(\)\s]', '', phone.strip()) return clean_phone if len(re.sub(r'[^\d]', '', clean_phone)) >= 10 else '' def _parse_amount(self, amount: Any) -> Optional[float]:"""Parse monetary amounts""" if amount is None: return None try: if isinstance(amount, (int, float)): return float(amount) elif isinstance(amount, str): # Remove currency symbols and commas clean_amount = re.sub(r'[^\d.]', '', amount) return float(clean_amount) if clean_amount else None except ValueError: return None return None def _generate_sam_url(self, notice_id: str) -> str:"""Generate SAM.gov URL for opportunity""" return f"https://sam.gov/opp/{notice_id}/view" if notice_id else"" # ... additional helper methods for resource links, organization types, etc. # This is just a fraction of the total parsing code needed! # Usage example (still complex after 400+ lines of parsing code) def process_sam_response(sam_response: Dict) -> List[Dict]:"""Process SAM.gov API response""" parser = SAMOpportunityParser() opportunities = [] for opp_data in sam_response.get('opportunitiesData', []): parsed_opp = parser.parse_opportunity(opp_data) opportunities.append(parsed_opp) return opportunities
This is just 60% of the required parsing code! Full production parsing includes:

Comparison: Clean Alternative API

GovCon API Response (No Parsing Required)

Here's the same opportunity data in a clean, flat structure:

{"data": [ {"notice_id":"abc123def456ghi789","title":"Software Development Services","solicitation_number":"W52P1J-25-R-0001","agency":"Department of Defense","department":"Department of Defense","sub_agency":"Army Contracting Command","office":"ACC - Detroit Arsenal","posted_date":"2025-11-01","response_deadline":"2025-11-30T17:00:00-05:00","notice_type":"Solicitation","set_aside_type":"Small Business Set-Aside","set_aside_code":"SBA","naics": ["541511","541512"],"naics_titles": ["Custom Computer Programming Services","Computer Systems Design Services"],"primary_naics":"541511","contact_name":"John Smith","contact_email":"[email protected]","contact_phone":"586-555-1234","secondary_contact":"Jane Doe","secondary_email":"","secondary_phone":"586-555-5678","performance_city":"Warren","performance_state":"MI","performance_state_name":"Michigan","performance_zip":"48397","performance_country":"USA","performance_address":"123 Main Street, Suite 100","sam_url":"https://sam.gov/opp/abc123def456ghi789/view","description_text":"The Army requires software development services for...","award_date":"2025-12-15","award_number":"W52P1J-25-C-0001","award_amount": 150000.00,"awardee_name":"ACME Software Solutions","awardee_location":"Detroit, MI","awardee_uei":"ABC123DEF456","archive_date":"2025-12-16","last_updated":"2025-11-01T10:30:00Z","active": true } ],"pagination": {"total": 1247,"limit": 100,"offset": 0,"has_next": true } }

Simple Processing (5 Lines vs 400+ Lines)

import requests # Get clean, parsed data instantly response = requests.get( 'https://govconapi.com/api/v1/opportunities/search', headers={'Authorization': 'Bearer your_api_key'}, params={'naics': '541511', 'limit': 100} ) opportunities = response.json()['data'] # Process clean data directly - no parsing needed! for opp in opportunities: print(f"Title: {opp['title']}") print(f"Agency: {opp['agency']}") # Clean, not nested print(f"Contact: {opp['contact_email']}") # Direct access print(f"Description: {opp['description_text']}") # Included! print(f"Award Amount: ${opp['award_amount'] or 'TBD'}") # Integrated print("---") # That's it! No parsing complexity, no error handling, no data normalization.

Development Time Comparison

Illustrative estimates based on a mid-sized integration project. Your actual time will vary based on scope, team experience, and existing tooling.

Task SAM.gov API GovCon API Time Saved
Data Structure Analysis 8 hours 0 hours 8 hours
Parsing Code Development 40 hours 0 hours 40 hours
Error Handling 16 hours 2 hours 14 hours
Testing & Debugging 20 hours 4 hours 16 hours
Data Validation 12 hours 1 hour 11 hours
Documentation 8 hours 1 hour 7 hours
Maintenance (yearly) 40 hours 2 hours 38 hours

Illustrative Time Saved: ~134 hours (3.5 weeks of full-time development)

Example Cost Savings at $75/hour: ~$10,050 in the first year

Based on a senior developer rate and a typical mid-sized project. Your numbers will vary.

Why SAM.gov's Structure is So Complex

Historical Technical Debt

Government vs. Commercial API Design

Aspect Government APIs Commercial APIs
Design Priority Compliance & completeness Developer experience
Data Structure Preserves source formats Optimized for consumption
Field Naming Regulatory terminology Intuitive naming
Breaking Changes Rarely allowed Managed with versioning
Performance Secondary concern Primary design goal

Alternative: Pre-Parsed Data

If you'd rather skip the parsing work, APIs like GovCon API provide the same data in a flat structure:

View simplified API format →

Summary

SAM.gov's nested JSON structure reflects the complexity of federal contracting data. The parsing code above handles the main patterns you'll encounter. For simpler use cases, consider using pre-parsed data sources.

Related Guides

Official Resources

Community Feedback

Help improve this guide for fellow developers. Press Ctrl+Enter to share issues you encountered, missing information, or suggestions.